EA042775B1 - BI-SPECIFIC ANTIBODY CONSTRUCTS FOR CDH3 AND CD3 - Google Patents

Description

The present invention relates to a bispecific antibody construct comprising a first human binding domain that binds to human CDH3 on the surface of a target cell and a second binding domain that binds to human CD3 on the surface of a T cell. In addition, the present invention provides a polynucleotide encoding an antibody construct, a vector containing said polynucleotide, and a host cell transformed or transfected with said polynucleotide or vector. In addition, the present invention provides a method for producing an antibody construct of the invention, a medical use of said antibody construct, and a kit containing said antibody construct.

Introduction

The cadherin superfamily spans over 100 members in humans, including the so-called classical cadherins P-cadherin, E-cadherin, N-cadherin, and R-cadherin, each of which is active in a separate set of tissues (Takeichi M. Development, 102:639-55( 1988), van Roy R, Nature Rev., V14:121-134 (2014)). Cadherins, in particular, play an important role in adult tissue and organ development, as well as in the homeostasis of various tissues (Conacci-Sorrell M., et al., J Clin Invest, 109:987-91, (2002)). Cadherins are transmembrane glycoproteins that regulate cell-cell adhesion processes by forming calcium-dependent bonds and by converting mechanical stimuli into electrochemical activity, a process called mechanotransduction (Gumbiner J. Cell. Biol., 148:399-404 (2000); Yagi, et al., Genes Dev., 14:1169-1180 (2000, Parades et al. Biochimica et Biophysica Acta 1826, 297-311 (2012)).

Placental cadherin (P-cadherin), also known as calcium-dependent cell-to-cell adhesion protein 3 (CDH3), is a 118 kDa protein with a large extracellular domain (ECD) of about 800 amino acids. The CDH3 ECD contains five cadherin repeats, referred to herein as (extracellular) domains 1-5/Dom1-Dom5/D1-D5, each containing about 110 amino acids. The proteins with the highest sequence homology to P-cadherin are Ecadherin with 53% sequence homology and N-cadherin with 39% homology. The expression level of P-cadherin is considered low in healthy adults and is limited to the basal or lower layers of stratified epithelium, including prostate and skin, breast myoepithelial cells (Takeichi M. J Cell Biol 103:2649-58, (1986) and Shimoyama Y, et al., Cancer Res., 49:212833 (1989)). While not lethal in gene knockout mice, loss of Pcadherin function has been shown to be associated with developmental defects as well as hyperplasia and dysplasia of the mammary epithelium. (G. L. Radice et al. J. Cell Biol. 139: 1025-1032 (1997)).

In contrast to the low levels of gene expression in healthy individuals, P-cadherin expression is upregulated in the context of several diseases, including, for example, immune diseases such as Crohn's disease and colitis (Hardy, et al., Gut 50:513-519 (2002)) . In addition, P-cadherin is believed to play an important role in the invasive nature of cancer cells (Furukawa, et al., Microscopy Res. Technique 38(4):343-352 (1997), Parades et al. Clin Cancer Res. 11 (16), 5869-5877 (2005), Parades et al Biochimica et Biophysica Acta 1826, 297-311 (2012)). An increase in P-cadherin synthesis has been described in a variety of tumors including colorectal, lung (NSCLC), breast (triple negative), pancreas, head and neck, thyroid, cervix, ovary, and stomach (Milic et al. Cancer Res 68: (19) 7760-7768 (2008), Imai et al Clin Cancer Res 14(20) 6487-6495 (2008), Paredes et al Clin Cancer Res 11 (16) 5869-5877 (2005), Dasgupta et al. Oral Oncology 42, 306-316 (2006), Jarzab et al Cancer Res 65: (4) 1587-1597 (2005), Patel et al Int J Cancer: 106, 172-177 (2003), Kim et al Human Pathology 41, 877-885 (2010)). In addition, increased expression of P-cadherin has been found to correlate with poor patient survival in various types of cancer (Sun L et al. Am J Pathol. 2011; 179:380-90; Gamallo, Modern Pathology, 14:650-654, (2001); Stefansson, et al., J. Clin. Oncol. 22(7): 1242-1252 (2004), Parades et al. Cancer Res. 64, 8309-8317 (2004)), Taniuchi K et al. Cancer Res. 2005; 65:3092-9; Paredes J et al. Clinic Cancer Res. 2005; 11:5869-77; Hardy R.G. et al. gut. 2002; 50:513-9; Peralta Soler A et al. cancer. 1999; 86:1263-72).

After cardiovascular disease, neoplasms are the second leading cause of death in the category of non-communicable diseases, resulting in approximately 8.3 million deaths worldwide in 2013 (GBD 2013 Lancet 2015; 385: 117-71). Since 1990, the absolute number of cancer cases has increased by 45.6%, due to the fact that the world population is growing, that people are getting older and that the prevalence of established risk factors (e.g. smoking, overweight, physical inactivity) is also increasing (GBD 2013 Lancet 2015; 385: 117-71, Torre LA et al CA Cancer J Clin 2015; 65:87-108). In men, lung cancer is the leading cause of death, with an estimated number of about 1.1 million cases per year, followed by liver and stomach cancers. In women, breast cancer is the leading cause of death with about 0.5 million cases per year, followed by lung and colon cancer. An estimated 14 million new cancer cases in 2012 worldwide, led by lung cancer with ~1.2 million in men and breast cancer with ~1.7 million in women (Torre LA et al. CA Cancer J Clin 2015; 65:87-108). Comprehensive cancer represents a major burden on society worldwide and there is a high need for treatment options to combat this disease.

- 1 042775

Common cancer treatment strategies include surgery followed by chemotherapy, radiation therapy, or, more recently, targeted therapy, or combinations of both (eg, NCCN guidelines for oncology). Chemotherapeutic agents include nucleotide analogs such as 5-fluororacil (5-FU), DNA damaging agents such as oxaliplatin inhibitors, and topoisomerase inhibitors such as irinotecan or microtubules such as docetaxel, all of which result in inhibition of tumor cell proliferation. Targeted therapies include, for example, small molecules that selectively inhibit mutant oncogenic kinases, such as the selective BRAF V600E compound vemurafenib (Garbe C. et al., Recent Results Cancer Res. 2014; 201:215-25), which is approved for the treatment of melanoma or the anaplastic lymphoma kinase (ALK) inhibitors crizotinib and ceritinib approved for the treatment of lung cancer (Pall G. Curr Opin Oncol. 2015; 27:118-24). In addition, there are antibody-based compounds such as cetuximab and panitumumab that recognize and inactivate the epidermal growth factor receptor (EGFR) and are approved for the treatment of KRAS wild-type colon cancer (Tol J, Punt CJ. Clin Ther. 2010; 32 :437-53) or trastuzumab, which recognizes Her2 and is approved for the treatment of breast cancer (Ahmed S. et al. Breast Cancer. 2015; 22:101-16).

Despite the development of a wide range of therapeutic interventions, except for sporadic cases, there is no cure for this heterogeneous disease. The heterogeneity of individual tumors leads to the fact that only a limited number of patients respond to a particular therapy and, in addition, resistance to therapeutic agents develops during tumor progression (Jamal-Hanjani M. et al. Clin Cancer Res. 2015; 21:1258-1266 ). Finally, drug side effects can also lead to interruption or discontinuation of treatment, although newly developed targeted therapies are expected to be better tolerated (http://www.cancer.net/navigating-cancercare/how-cancer-treated/chemotherapy/side -effects-chemotherapy; Sun GC et al. Anticancer Agents Med Chem. 2015 Mar 17 [Online before print release]).

In addition, for these reasons, there is still a high medical need to develop new drugs.

CDH3 overexpression in tumor cells provides the basis for a new approach to cancer treatment using a bispecific antibody that recognizes CDH3 overexpressing tumor cells and kills them by redirecting cytotoxic T cells that are recognized by the CD3 binding portion of this antibody. Expression analyzes, which have been performed at either the RNA or protein level, using primarily immunohistochemistry, demonstrate that CDH3 is expressed at a high and well-detectable level in a wide range of different types of cancer, including colon cancer (Milic et al., Kita et al., Imai et al., http://www.proteinatlas.org/), lung, including non-small cell and small cell lung cancer (Imai et al., human protein atlas), breast, preferably triple negative cancer (Perou et al., Paredes et al., Turashvili et al., Imai et al.), pancreas (Imai et al., Taniuchi et al.), head and neck, including tongue squamous cell carcinoma (Dasgupta et al; own unpublished data), thyroid (Jarzab et al., Rocha et al., human protein atlas), cervix (Imai et al., Han et al.), ovary, preferably stage II and above (Patel et al human protein atlas), stomach (Kim et al; Imai et al), endometrium (Sugiyama Y et al. Clin Cancer Res. 2003; 9:5589-600; human protein atlas), cholangiocarcinoma (Baek S et al. Anat Cell Biol. 2010; 43:110-7; Imai et al.), bladder (Imai et al., own unpublished data), prostate (Imai et al.; own unpublished data), testis (Imai et al), soft tissue sarcoma (Imai et al), esophagus (own unpublished data), kidney (own unpublished data).

Future analyzes may show increased expression in additional tumors or subtypes, which may then also become relevant for such therapy. Fotouhi et al., for example, have demonstrated a decrease in CDH3 promoter methylation in small intestinal neuroendocrine tumors, which can lead to increased CDH3 protein expression (Fotouhi O et al. Epigenetics. 2014; 9:987-97). A positive correlation has been observed between increased expression of CDH3 and miR205, a miRNA that appears to be a specific marker for lung squamous cell carcinoma (Huang W et al. Chin Med J (Engl). 2014; 127:272-8).

In contrast to the increased expression of CDH3 in tumor tissue, its expression in normal tissue is low or not detected (Milic et al., Imai et al., Taniuchi et al., Rocha et al., Dasgupta et al., Han et al., Patel ., Kim et al., Sugiyma et al., Jarzab et al., human protein atlas). Taken together, this suggests that CDH3 is a suitable target for the proposed bispecific antibody approach.

Many antibodies have been described that bind to P-cadherin, and the most thoroughly characterized have been published in the following patents: 2010126137; WO 2010054007, WO 2011056997, WO 2011080796, WO 2011071541.

Some of the antibodies known in the art block the function of CDH3, for example by interfering with its adhesive properties, such as the anti-CDH3 antibody PF-03732010 (Zhang CC et al.

- 2 042775

Clinic Cancer Res. 2010; 16:5177-88). In vitro, this antibody results in disruption of 3D-cnepougoe with changes in intracellular signaling such as dissociation of β-catenin at nanomolar concentrations, but does not inhibit proliferation. In vivo inhibition of tumor growth and metastasis formation and long-term survival have been observed in preclinical models at doses of 10 mg/kg and in a manner that depends on the level of CDH3 expression. It is mechanically possible that in vivo antitumor effects are further mediated by the process of antibody-dependent cellular cytotoxicity (ADCC), which involves cell killing by binding of cytotoxic T cells to the Fc domain of immunoglobulins. This mechanism of action has been attributed to some of the claimed anti-CDH3 antibodies. Although blocking P-cadherin activity has been described as an approach to inhibiting the growth of tumors expressing P-cadherin, the results achieved are often unsatisfactory because they delay tumor growth or, at best, induce tumor stasis but do not eliminate the tumor.

In order to improve their inhibitory effect on tumor growth, antibodies can be conjugated to cytotoxic or cytostatic agents (e.g., a chemotherapeutic agent, toxin, radioactive isotope, or the like), resulting in the formation of antibody drug conjugates (ADCs) or immunoconjugates in general. Immunoconjugates allow targeted delivery of the drug fragment to the tumor and their intracellular accumulation. The effectiveness of such immunoconjugates in killing tumor cells is highly dependent on various parameters such as the internalization behavior of the target molecule, which is usually found in the cell membrane, with an extracellular domain, as well as the mechanism of action of the respective immunoconjugate. In general, binding to an immunoconjugate has been shown to significantly increase the antitumor efficacy of various antibodies.

Recently, bispecific molecules that bind to the target molecule as well as to T cells have shown promising results, circumventing many of the aforementioned shortcomings. These drug molecules do not rely on complete blocking of the target function or on the state of proliferation of the target cell, as they exploit the very effective natural effect of killing T cells. One example of such bispecific molecules are the bispecific anti-target x anti-CD3 single chain antibody constructs, which have previously been shown to mediate T cell-mediated killing of target cells with very high efficiency, see eg blinatumomab.

To distinguish a bispecific scFv molecule from a good drug that will ultimately help cure patients, many different criteria must be met. The combination and severity of these criteria make the appearance of molecules that satisfy these needs an extremely rare and unpredictable event. The most obvious criterion is the efficacy of drug candidates.

The effectiveness of such molecules depends on a variety of parameters on the target binding domain as well as on the T-cell recruiting portion of the bispecific molecule. Especially on the target binding moiety of the bispecific scFv molecule are unpredictable as they vary greatly depending on the nature of the target molecule, including its tertiary as well as quaternary structure. Among the performance-determining parameters, the binding kinetics between the target and the bispecific binding molecule, as well as the precise binding region, the so-called binding epitope, play an important role.

Anti-CDH3 monoclonal antibodies - generally true of any other monoclonal antibodies - function by highly specific recognition of their target molecules. They recognize only one site or epitope on their target CDH3 molecule. In addition, many antibodies have been found to function in a species-specific manner. However, this species specificity, inherent not only to CDH3 monoclonal antibodies (and their fragments), but also to monoclonal antibodies in general, is a significant obstacle to their development as therapeutic agents for the treatment of human diseases. To be approved in the marketplace, any new drug candidate must undergo rigorous testing.

This trial is divided into preclinical and clinical phases; while the latter is performed on human patients, the former on animals. The purpose of a preclinical trial is to prove that a drug candidate has the desired activity and, most importantly, is safe. Only when the animal safety and possible efficacy of a drug candidate have been established in preclinical testing is that drug candidate likely to be approved for human clinical trial by the appropriate regulatory authority. Candidate drugs can be tested for animal safety in three ways: (i) in relevant species, i.e. species in which candidate drugs can recognize orthogenic antigens, (ii) in a transgenic animal containing human antigens, and (iii) ) using a drug candidate surrogate that can bind orthogenic antigens present in the animal. The limitations of transgenic animals is that this technology is usually limited to rodents.

- 3 042775

There are significant differences in physiology between rodents and humans, and safety results cannot easily be extrapolated to humans. The limitations of a surrogate for a candidate drug are a different composition of the substance compared to the actual drug candidate, and rodents with the limitation discussed above are frequently used animals. Therefore, preclinical data from rodents have limited predictive power for drug candidate.

The approach of choice for safety testing is to use an appropriate species, preferably a primate, and because of genetic similarity, a chimpanzee. However, chimpanzees are considered an endangered species, and due to their human-like nature, the use of such animals for drug safety testing has been banned in Europe and severely restricted elsewhere. Although the T-cell bispecific single chain antibodies described in the art have great therapeutic potential for the treatment of cancer, most of these bispecific molecules are limited in that they are species-specific and recognize only the human antigen and probably a primate counterpart, .e. toque. Moreover, most of these bispecific molecules are further limited in that they cannot perform their desired function across species boundaries such that they can recognize human and primate homologues but cannot exert, for example, T-cell mediated cytotoxicity.

The essence of the invention

Since there is still a need for additional treatment options for the various types of cancer described herein, this document provides tools and methods to address this problem in the form of a bispecific antibody construct with a single binding domain targeting CDH3 and a second binding domain targeting CDH3. on CD3 on T cells.

The present invention provides, in a first aspect, a bispecific antibody construct comprising a first (preferably human) binding domain that binds to a human CDH3 epitope cluster on the surface of a target cell and comprising a second (preferably human) binding domain that binds to human CD3 on surface of the T cell, wherein the human CDH3 epitope cassette is located at amino acid positions 291-363 (SEQ ID NO: 36) of human CDH3. In a preferred embodiment of the invention, the bispecific antibody construct is characterized in that the first binding domain also binds to macaque CDH3, preferably Macaca fascicularis CDH3.

This cross-species functionality means that the same molecule can be used in preclinical animal studies as well as human clinical studies. This results in highly comparable results and greatly increased predictive power in animal studies compared to species-specific surrogate molecules. Since both the CD3 and the CDH3 binding domain of the CDH3xCD3 bispecific antibody constructs of the invention are cross-species-specific and functional, i.e. reactive with human and macaque antigens, with a comparable effect of T-cell mediated cytotoxicity, it can be used both for preclinical evaluation of the safety, activity and/or pharmacokinetic profile of these binding domains in primates, and in identical form, as a drug in humans. It is understood that in a preferred embodiment of the invention, the cross-species specificity of the first and second binding domain of the antibody constructs of the invention is identical.

In view of the foregoing, the need to construct a surrogate construct of a CDH3xCD3 bispecific antibody for testing in a phylogenetically distant (from human) species disappears. As a result, an identical molecule can be used in preclinical animal studies, as intended for administration to humans in clinical trials, and for further market approval and administration of a therapeutic drug. The ability to use the same molecule for preclinical animal testing as for subsequent administration to humans virtually eliminates or at least significantly reduces the risk that data obtained from preclinical animal testing have limited applicability to the human case. In short, obtaining preclinical safety data in animals using the same molecule that will actually be administered to humans goes a long way towards ensuring that the data is applicable to the appropriate scenario in humans. In contrast, in conventional approaches using surrogate molecules, said surrogate molecules must be molecularly adapted to the animal test system used for preclinical safety assessment. Thus, the molecule to be used in human therapy actually differs in sequence and probably in structure from the surrogate molecule used in preclinical testing in terms of pharmacokinetic parameters and/or biological activity, as a result of which the data obtained in preclinical testing on animals, have limited applicability and portability to human cases. The use of surrogate molecules requires the design, manufacture, purification and evaluation of an entirely new design. This requires additional costs for

- 4 042775 development and time needed to obtain this molecule. In sum, surrogates must be developed separately in addition to the actual drug to be used in human therapy, so two lines of development for two molecules need to be done. Therefore, the main advantage of the preferred human CDH3xCD3 bispecific antibody construct of the invention, exhibiting the cross-species specificity and functionality (i.e., reactivity) described herein, is that the identical molecule can be used for therapeutic agents in humans and for preclinical testing in animals. .

When using the cross-species-specific design of CDH3xCD3 bispecific antibodies according to the invention, it is also not required to adapt the test animal to a drug candidate intended for administration to humans, for example, to create transgenic animals. The CDH3xCD3 bispecific antibody construct (preferred for humans) of the invention, exhibiting cross-species specificity and reactivity according to the uses and methods of the invention, can be directly applied to preclinical testing in primates other than chimpanzees, such as macaques, without any genetic manipulation of animals. As is well known to those skilled in the art, approaches in which the test animal is adapted to the drug candidate always carry the risk that the results obtained in a pre-clinical safety trial are less representative and predictive of the human due to modification of the animal. For example, in transgenic animals, the proteins encoded by the transgenes are often highly overexpressed. Thus, data obtained for the biological activity of an antibody against this protein antigen may be limited in its predictive value in humans, in which the protein is expressed at much lower, more physiological levels.

Another advantage of the preferably human construct of the CDH3xCD3 bispecific antibody of the invention is the ability to extract multiple blood samples when used as part of a preclinical animal study, eg, in animal pharmacokinetic studies. Multiple blood extracts can be obtained much more easily with non-chimpanzee primates than with more primitive animals such as the mouse. Retrieval of multiple blood samples allows continuous testing of blood parameters to determine the biological effects induced by the CDH3xCD3 bispecific antibody construct of the invention. In addition, the extraction of multiple blood samples allows the investigator to evaluate the pharmacokinetic profile of the CDH3xCD3 (preferably human) bispecific antibody construct of the invention as defined herein. In addition, the potential side effects that may be induced by said CDH3xCD3 bispecific antibody construct of the invention reflected in blood parameters can be measured in different blood samples extracted during the administration of said antibody.

The advantages of the CDH3xCD3 bispecific antibody construct of the invention, as defined herein, demonstrating cross-species specificity, can be briefly summarized as follows:

First, the CDH3xCD3 bispecific antibody construct of the invention, as defined herein, used for preclinical testing is the same as in human therapy. Thus, it is not necessary to develop two independent molecules that may differ in their pharmacokinetic properties and biological activity. This is very important in that, for example, pharmacokinetic results are more directly transferable and applicable to human conditions than, for example, in conventional surrogate approaches.

Second, the use of the CDH3xCD3 bispecific antibody construct of the invention, as defined herein, to produce therapeutic agents in humans is less costly and labor intensive than surrogate approaches.

Third, the CDH3xCD3 bispecific antibody construct of the invention, as defined herein, can be used for preclinical testing not only in one primate species, but in a number of different primate species, thus limiting the risk of potential differences between primates and humans.

Fourth, the chimpanzee, as an endangered species for animal testing, can be avoided if necessary.

Fifth, multiple blood samples can be extracted for extensive pharmacokinetic studies.

Sixth, due to the human origin of the antibody constructs in accordance with the preferred embodiment of the invention, the development of an immune reaction against these binding molecules is minimal when administered to human patients. Induction of an immune response with antibodies specific for a drug candidate derived from a non-human species, eg mouse, leading to the development of human anti-mouse antibodies (HAMA) to therapeutic molecules of mouse origin.

- 5 042775

The therapeutic application of the CDH3xCD3 bispecific antibody construct of the invention provides a novel and inventive therapeutic approach to cancer, preferably solid tumors, more preferably carcinomas and other indications of cancer as indicated below, as shown in the following examples, the CDH3xCD3 bispecific antibody construct of the invention provides an advantage as a tool to kill CDH3-expressing human cancer cells. In addition, the cytotoxic activity of the CDH3xCD3 bispecific antibody constructs of the invention provides a preferred mode of action over classical (monospecific) IgG or ADC molecules.

To characterize drug candidates in animal models, species-to-species cross-reactivity is required, and the interspecies affinity gap must be maintained at a 10-fold affinity difference to provide a predictive animal model. Monkey macaques (and especially cynomolgus macaques) are considered to be among the most significant species for efficacy and toxicity testing with the highest predictive value. In order to analyze whether the antibody criteria meet the performance criteria, appropriate cytotoxicity assays need to be developed and qualified.

To prevent adverse side effects, it is necessary to ensure that the target binding portion of the bispecific antibody construct specifically binds to P-cadherin and not to one of its closest homologues described above or other proteins present in the body.

The antibodies of the invention preferably do not bind to the D1 extracellular domain of CHD3 (positions 108-215 of SEQ ID NO: 1), they preferably do not bind to the D4 extracellular domain of CHD3 (positions 441-546 of SEQ ID NO: 1), and they preferably do not bind with the extracellular domain D5 of CHD3 (positions 547-650 of SEQ ID NO: 1).

According to the most preferred embodiment of the present invention, the bispecific antibody construct is characterized in that the first binding domain binds to an epitope that is located at amino acid positions 291-327 (SEQ ID NO: 34) of human CDH3. Antibodies that specifically bind to amino acid positions 291-327 (SEQ ID NO: 34) preferably do not bind to the extracellular D3 domain of human CHD3 CDH3 (positions 328-440 of SEQ ID NO: 1).

A preferred antibody construct according to the invention can also be defined as a bispecific antibody construct comprising a first binding domain (preferably human) that binds to a human CDH3 epitope on the target cell surface and a second binding domain (preferably human) that binds to human CD3 on the surface of the T cell, wherein the antibody construct binds to the same epitope as or competes for CDH3 binding with an antibody designated CDH3-11, CDH3-12, CDH3-13 or CDH3-14, i.e. an antibody containing:

VH region as shown in SEQ ID NO: 155 and VL region as shown in SEQ ID NO: 156, VH region as shown in SEQ ID NO: 165 and VL region as shown in SEQ ID NO: 166, VH region , as shown in SEQ ID NO: 175 and the VL plot, as shown in SEQ ID NO: 176; or a VH region as shown in SEQ ID NO: 185 and a VL region as shown in SEQ ID NO: 186.

In another aspect of the present invention, the bispecific antibody construct is characterized in that the first binding domain binds to an epitope that is located at amino acid positions 328-363 (SEQ ID NO: 35) of human CDH3.

A preferred antibody construct according to the invention can also be defined as a bispecific antibody construct comprising a first binding domain (preferably human) that binds to a human CDH3 epitope on the target cell surface and a second binding domain (preferably human) that binds to human CD3 on the surface of the T cell, wherein the antibody construct binds to the same epitope as or competes for binding to CDH3 with an antibody designated CDH3-24, i.e. an antibody containing a VH region as shown in SEQ ID NO: 285 and a VL region, as shown in SEQ ID NO: 286.

Another preferred antibody construct according to the invention can also be defined as a bispecific antibody construct comprising a first binding domain (preferably human) that binds to a human CDH3 epitope on the target cell surface and a second binding domain (preferably human) that binds to human CD3 on T cell surface, wherein the antibody construct binds to the same epitope as or competes for CDH3 binding with an antibody designated CDH3-25, CDH3-26, or CDH3-27, i.e., an antibody containing a VH region as shown in SEQ ID NO: 295 and VL plot as shown in SEQ ID NO: 296, VH plot as shown in SEQ ID NO: 305 and VL plot as shown in SEQ ID NO: 306; or a VH region as shown in SEQ ID NO: 315 and a VL region as shown in SEQ ID NO: 316.

Whether or not an antibody construct competes for binding with another

- 6 042775 antibody construct, in a competitive assay such as a competitive ELISA (enzyme-linked immunosorbent assay) or a cellular competitive assay. Avidin-bound microparticles (beads) can also be used. Like an avidin-coated ELISA plate, when interacting with a biotinylated protein, each of these beads can be used as a substrate on which the assay can be performed. The antigen is applied to the bead and then pre-coated with the first antibody. A second antibody is added and any additional binding determined. Reading is done by flow cytometry.

In another embodiment of the invention, the bispecific antibody construct is characterized in that the first binding domain binds to an epitope located at amino acid positions 328-363 (SEQ ID NO: 35) of human CDH3 and to an epitope located at amino acid positions 404-440 ( SEQ ID NO: 390) human CDH3.

In addition, the antibody constructs of this invention preferably do not bind to an epitope that is at amino acid positions 216-252 or 253-290 of human CDH3 as shown in SEQ ID NO: 1.

In addition, the antibody constructs of this invention preferably do not bind to an epitope that is at amino acid positions 364-403 of human CDH3 as shown in SEQ ID NO: 1.

Antibody constructs that specifically bind to an epitope that is at amino acid positions 328-363 (SEQ ID NO: 35) preferably do not bind to the extracellular D2 domain of human CHD3 CDH3 (positions 216-327 of SEQ ID NO: 1).

One of the advantages of the present invention is to provide a bispecific antibody construct containing a binding domain that binds to CD3 and a binding domain capable of binding to CDH3, while both binding domains show cross-species specificity for human and macaque CDH3. Surprisingly, the CDH3xCD3 bispecific antibody constructs of the invention not only specifically bind to human and macaque CDH3 and CDH3 homologues, but also exhibit T-cell mediated cytotoxicity in human and macaque CDH3 assay systems. Advantageously, the invention provides CDH3xCD3 bispecific antibody constructs that exhibit T-cell mediated cytotoxicity in humans and macaques. This advantage is achieved by CDH3xCD3 bispecific antibody constructs that bind to an epitope cassette located at amino acid positions 291-363 (SeQ ID NO: 36) of human CDH3.

It should be noted that the singular forms include references to the plural unless the context clearly dictates otherwise. Thus, for example, a reference to a reagent includes one or more of these various reagents, and a reference to a method includes a reference to equivalent steps and methods known to those skilled in the art, which may be modified or replaced by the methods described herein.

Unless otherwise indicated, a term at least preceding a series of elements should be considered to refer to each element in the series. Those skilled in the art will understand or be able to determine many equivalents to the specific embodiments of the invention described herein using no more than standard experimentation. Such equivalents are covered by the present invention.

The term and/or wherever it is used in this document, includes the meaning of and, or and all or any other combination of elements associated with the specified term. As used herein, about or approximately means within ±20%, preferably within ±15%, more preferably within ±10%, and most preferably within ±5% of a given value or range.

In this description and claims, unless the context otherwise requires, the word contains and its variations, such as includes and containing, should be understood as including the specified integer value or step, or group of integer values or steps, but not excluding any other integer value or a stage or group of integers or stages. As used herein, the term containing may be replaced by the term which contains or including, or sometimes, when used herein, the term having.

In this context, the term consisting of excludes any element, step or ingredient not listed in the claim element. In this context, the term consisting essentially of does not exclude materials or steps that do not significantly affect the essential and novel features of the claimed invention.

In each case in this document, any of the terms including, consisting essentially of, and consisting of may be replaced by any of the other two terms.

The term antibody design refers to a molecule in which the structure and/or function is(s) based on the structure and/or function of an antibody, eg, a full length or whole immunoglobulin molecule. Thus, the antibody construct is able to bind to its specific target.

- 7 042775 or antigen. In addition, the design of the antibody in accordance with the invention contains the minimum structural requirements for the antibody, which allow binding to the target. This minimum requirement may, for example, be defined by having at least three light chain CDRs (i.e., CDR1, CDR2, and CDR3 of the VL region) and/or three CDRs of the heavy chain (i.e., CDR1, CDR2, and CDR3 of the VH region), preferably from all six CDRs. Antibodies on which the constructs of the invention are based include, for example, monoclonal, recombinant, chimeric, deimmunized, humanized and human antibodies.

In the definition of antibody constructs according to the invention, full-length or whole antibodies, including camel antibodies and other immunoglobulin antibodies, obtained by methods or processes of biotechnological or protein engineering are provided. These full length antibodies can be, for example, monoclonal, recombinant, chimeric, deimmunized, humanized and human antibodies. Also included in the definition of antibody constructs are full length antibody fragments such as VH, VHH, VL, (s) dAb, Fv, Fd, Fab, Fab', F(ab') ₂ or r IgG, (half antibody). Antibody constructs according to the invention can also be modified antibody fragments, also referred to as antibody variants, such as scFv, di-scFv or bi(s)-scFv, scFv-Fc, scFv-zip, scFab, Fab2, Fab3, diabodies, single chain diabodies, tandem diabodies (Tandab), tandem di-scFvs, tandem tri-scFvs, mini-antibodies exemplified by a structure that looks like this: (VH-VL-CH3) ₂ , (scFvCH3) ₂ , ((scFv ) ₂ -CH3 + CH3), ((scFv) ₂ -CH3) or (scFv-CH3-scFv) ₂ , multibodies such as triabodies or tetrabodies, and single domain antibodies such as nanoparticles or single variable domain antibodies containing only one variable domain, which may be a VHH, VH, or VL, that specifically binds an antigen or epitope independently of other V regions or domains.

In addition, the definition of the term antibody construct includes monovalent, divalent, and polyvalent/multivalent constructs, and thus monospecific constructs that specifically bind to only one antigenic construct, as well as bispecific and polyspecific/multispecific constructs that specifically bind more than one antigenic construct. structure, for example, two, three or more, through different binding domains. Moreover, the definition of the term antibody construct includes molecules consisting of only one polypeptide chain, as well as molecules consisting of more than one polypeptide chain, and the chains can be either identical (homodimers, homotrimers, or homogeneous oligomers) or different (heterodimer, heterotrimer or heterooligomer). Examples of the above antibodies and variants or derivatives thereof are described in particular in Harlow and Lane, Antibodies a laboratory manual, CSHL Press (1988) and Using Antibodies: a laboratory manual, CSHL Press (1999), Kontermann and Dubel, Antibody Engineering, Springer , 2nd ed. 2010 and Little, Recombinant Antibodies for Immunotherapy, Cambridge University Press 2009.

The antibody constructs of this invention are preferably in vitro antibody constructs. The term refers to an antibody construct as defined above, in which all or part of the variable region (e.g., at least one CDR) is generated in non-immune cell selection, e.g., in vitro phage display, protein chip, or any other method, wherein candidate sequences can be tested for their ability to bind to an antigen. Thus, the term preferably excludes sequences generated solely by genomic rearrangement in an immune cell in an animal. A recombinant antibody is an antibody obtained using recombinant DNA technology or genetic engineering.

In this context, the term monoclonal antibody (mAb) or monoclonal antibody construct refers to an antibody derived from a population of essentially homogeneous antibodies, i.e., from a population in which individual antibodies are identical, except for possible natural mutations and/or post-translational modifications. (eg isomerization, amidation), which may be present in minor amounts. Monoclonal antibodies are highly specific, which are directed against a single antigenic site or antigen determinant, in contrast to conventional (polyclonal) antibodies, which usually include different antibodies directed against different determinants (or epitopes). In addition to their specificity, monoclonal antibodies have the advantage that they are synthesized by the culture of the hybridoma, therefore uncontaminated by other immunoglobulins. The modifier monoclonal indicates that the antibody is derived from a substantially homogeneous population of antibodies; it should not be interpreted as requiring the production of an antibody by any particular method.

To obtain monoclonal antibodies, you can use any technique that provides antibodies obtained by continuous cultures of cell lines. For example, the monoclonal antibodies to be used can be made using the hybridoma method first described by Koehler et al., Nature, 256: 495 (1975) or can be made by recombinant DNA techniques (see, for example, US Pat. No. 4,816,567). Examples of additional methods for preparing human monoclonal antibodies include the trioma technique, the human B-cell hybridoma technique (Kozbor,

- 8 042775

Immunology Today 4 (1983), 72) and the EBV hybridoma technique (Cole et al., Monoclonal Antibodies and Cancer Therapy, Alan R. Liss, Inc. (1985), 77-96).

Hybridomas are then screened using standard methods such as enzyme immunoassay (ELISA) and surface plasmon resonance (BIACORE™) to identify one or more hybridomas that produce an antibody that specifically binds to said antigen. As an immunogen, you can use any form of the corresponding antigen, for example, a recombinant antigen, natural forms, any of its variants or fragments, as well as their antigenic peptide. Surface plasmon resonance used in the BIAcore system can be used to increase the efficiency of phage antibodies that bind to a target antigen epitope such as CDH3 or CD3 epsilon (Schier, Human Antibodies Hybridomas 7 (1996), 97-105; Malmborg, J. Immunol Methods 183 (1995), 7-13).

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

In addition to using display libraries, an appropriate antigen can be used to immunize a non-human animal, such as a rodent (such as a mouse, hamster, rabbit, or rat). In one embodiment, the non-human animal includes at least a portion of the human immunoglobulin gene. For example, mouse strains deficient in the production of mouse antibodies can be engineered with large fragments of human Ig loci. Using hybridoma technology, it is possible to produce and select antigen-specific monoclonal antibodies derived from genes with the desired specificity. See, for example, XENOMOUSE™, Green et al. (1994) Nature Genetics 7:13-21, US 2003-0070185, WO 96/34096 and WO 96/33735.

A monoclonal antibody can also be obtained from a non-human animal and then modified, eg, humanized, deimmunized, rendered chimeric, etc., using recombinant DNA techniques known in the art. Examples of modified antibody constructs include humanized antibody variants, non-human origin, affinity matured antibodies (see, e.g., Hawkins et al. J. Mol. Biol. 254, 889896 (1992) and Lowman et al., Biochemistry 30, 10832-10837 ( 1991)) and antibody mutants with altered effector function(s) (see, for example, US Pat. No. 5,648,260, Kontermann and Dubel (2010), loc. cit. and Little (2009), loc. cit).

In immunology, affinity maturation is the process by which B cells produce antibodies with increased affinity for an antigen during an immune response. Upon repeated exposure to the same antigen, the host will produce antibodies with consistently higher affinity. Like the natural prototype, in vitro affinity maturation is based on the principles of mutation and selection. In vitro affinity maturation has been successfully used to optimize antibodies, antibody constructs, and antibody fragments. Random mutations within the CDR are introduced using radiation, chemical mutagens, or error-correcting PCR. In addition, genetic diversity can be increased by strand shuffling. Two or three rounds of mutation and selection using display techniques such as phage display typically results in antibody fragments with low nanomolecular affinities.

A preferred type of amino acid substitution variant of antibody constructs involves substitution of one or more hypervariable region residues of the parent antibody (eg, humanized or human antibody). Typically, the resulting variant(s) selected for further modification will have improved biological properties relative to the original antibody from which they are derived. A suitable method for obtaining such substitution variants involves affinity maturation using phage display. Briefly, several hypervariable region sites (eg, 6-7 sites) are subjected to mutagenesis to obtain all possible amino acid substitutions at each site. The antibody variants thus obtained are then detected in monovalent form on the surface of the filamentous bacteriophage particles in the form of their fusion with the bacteriophage M13 gene III product packaged in each particle. The phage display variants are then screened for their biological activity (eg, binding affinity) as described herein. In order to identify possible sites for modification of the hypervariable regions, alanine-scanning mutagenesis can be performed to identify hypervariable region residues that provide significant antigen binding. Alternatively or additionally, it may be useful to analyze the crystal structure of the antigen-antibody complex in order to identify points of contact between the binding domain and, for example, human CDH3. Such contact residues and neighboring residues are potential targets for replacement, according to the methods presented in this document. Once such variants have been generated, they are screened as described herein,

- 9 042775 and antibodies with improved properties, as established in one or more relevant studies, are selected for further modification.

The monoclonal antibodies and antibody constructs of this invention specifically include chimeric antibodies (immunoglobulins) in which the heavy and/or light chain region is identical or homologous to the corresponding sequences of antibodies derived from a particular species or belonging to a particular class or subclass of antibodies, while the remainder of the chain ( chains) is identical or homologous to the corresponding sequences of antibodies derived from another species or belonging to another class or subclass of antibodies, as well as fragments of such antibodies, provided that they exhibit the desired biological activity (US Patent No. 4816567; Morrison et al., Proc. Natl Acad. Sci. USA, 81: 6851-6855 (1984)). Chimeric antibodies contemplated herein include primatized antibodies comprising variable domain antigen-binding sequences derived from a non-human primate antibody (eg, marmosets, great apes, etc.) and human constant region sequences. Various approaches to the creation of chimeric antibodies have been described. See, for example, Morrison et al., Proc. Natl. Acad. ScL U.S.A. 81:6851, 1985; Takeda et al., Nature 314:452, 1985, Cabilly et al., US Pat. No. 4,816,567; Boss et al., US patent No. 4816397; Tanaguchi et al., EP 0171496; EP 0173494 and GB 2177096.

An antibody, antibody construct, or antibody fragment can also be modified by specific deletion of human T cell epitopes (a technique called deimmunization) by methods described in, for example, WO 98/52976 or WO 00/34317. Briefly, the variable regions of the heavy and light chains of an antibody can be analyzed for peptides that bind to MHC (major histocompatibility complex) class II; these peptides are potential T cell epitopes (as defined in WO 98/52976 and WO 00/34317). A computer modeling approach called peptide threading can be applied to discover potential T-cell epitopes, and in addition the database of human MHC class II binding peptides can be examined for motifs present in VH and VL sequences as described in WO 98/ 52976 and WO 00/34317. These motifs bind to any of the 18 major MHC class II DR allotypes and thus constitute potential T-cell epitopes. Identified potential T cell epitopes can be eliminated by substitution of a small number of amino acid residues in the variable domains, or preferably by single amino acid substitutions. Usually, conservative substitutions are made. Often, but not exclusively, an amino acid common to a position in human germline antibody sequences can be used. Human germline sequences are disclosed in, for example, Tomlinson, et al. (1992) J. Mol. Biol. 227:776-798; Cook, G. P et al. (1995) Immunol. Today Vol. 16 (5): 237-242 and Tomlinson et al. (1995) EMBO J. 14:14:46284638. The VBASE catalog is a complete catalog of human immunoglobulin variable region sequences (compiled by Tomlinson LA et al., MRC Protein Modification Center, Cambridge, UK). These sequences can be used as a source of human sequence, for example for frameworks and CDRs. Human consensus frameworks can also be used, for example, as described in US Pat. No. 6,300,064.

Humanized antibodies, antibody constructs, or fragments thereof (such as Fv, Fab, Fab', F(ab')2 or other antibody antigen-binding subsequences) are antibodies or immunoglobulins, mostly human sequences, that contain (a) minimal sequence(s) ) derived from non-human immunoglobulin. In most cases, humanized antibodies are human (recipient antibody) immunoglobulins in which residues from the hypervariable region (also CDR) of the recipient are replaced by residues from the hypervariable region (donor antibody) of a non-human species (e.g., rodents), such as mice, rat, hamster or rabbit having the desired specificity, affinity and potency. In some instances, Fv framework region (FR) residues of a human immunoglobulin are replaced by corresponding non-human residues. In addition, the humanized antibodies used herein may also contain residues that are not found in either the recipient antibody or the donor antibody. These modifications are made to further purify and optimize antibody performance. The humanized antibody may also contain at least an immunoglobulin constant region (Fc) fragment, typically derived from a human immunoglobulin. For additional information, see Jones et al., Nature, 321: 522-525 (1986); Reichmann et al., Nature, 332: 323-329 (1988) and Presta, Curr. Op. Struct. Biol., 2: 593-596 (1992).

Humanized antibodies or fragments thereof can be generated by substituting Fv variable domain sequences that are not directly related to antigen binding with equivalent sequences from human Fv variable domains. Exemplary methods for preparing humanized antibodies or fragments thereof are provided by Morrison (1985) Science 229:1202-1207; Oi et al. (1986) BioTechniques 4:214 and US 5,585,089; US 5693761; US 5693762; US 5,859,205 and US 6,407,213. These methods include isolating, manipulating, and expressing nucleic acid sequences that encode all or part of the immunoglobulin Fv variable regions of at least one heavy or light chain. Such nucleic acids can be obtained from hybridoma,

- 10 042775 producing an antibody against a given target, as described above, as well as from other sources. The recombinant DNA encoding the humanized antibody can then be cloned into an appropriate expression vector.

Humanized antibodies can also be generated using transgenic animals such as mice that express human heavy and light chain genes but are unable to express endogenous mouse immunoglobulin heavy and light chain genes. Winter describes an exemplary CDR grafting method that can be used to prepare the humanized antibodies described herein (US Pat. No. 5,225,539). All of the CDRs of a particular human antibody may be replaced with at least a portion of a non-human CDR, or only some of the CDRs may be replaced with non-human CDRs. It is only necessary to change the number of CDRs needed to bind a humanized antibody to a given antigen.

A humanized antibody can be optimized by introducing conservative substitutions, consensus sequence substitutions, germline substitutions, and/or back mutations. Such altered immunoglobulin molecules can be prepared by any of several methods known in the art (e.g., Teng et al., Proc. Natl. Acad. Sci. U.S.A., 80: 73087312, 1983; Kozbor et al., Immunology Today, 4 : 7279, 1983; Olsson et al., Meth. Enzymol., 92: 3-16, 1982 and EP 239400).

The term human antibody, human antibody construct, and human binding domain includes antibodies, antibody constructs, and binding domains having antibody regions, such as variable and constant regions or domains, that substantially correspond to human germline immunoglobulin sequences known in the art, including, for example, those described by Kabat et al. (1991) (loc. cit.). Human antibodies, antibody constructs, or binding domains of the invention may contain amino acid residues not encoded by human germline immunoglobulin sequences (e.g., mutations introduced by in vitro random or site-directed mutagenesis or by in vivo somatic mutation), for example, in CDRs and in particular CDR3. Human antibodies, antibody constructs, or binding domains may have at least one, two, three, four, five, or more positions substituted with an amino acid residue that is not encoded by a human germline immunoglobulin sequence. The definition of human antibodies, antibody constructs, and binding domains used herein also contemplates fully human antibodies that contain only non-artificial and/or genetically altered human antibody sequences as they can be generated using technologies or systems such as Xenomouse.

In some embodiments, the antibody constructs of the invention are isolated or substantially pure antibody constructs. Isolated or substantially pure, when used to describe an antibody construct disclosed herein, means an antibody construct that has been identified, separated, and/or isolated from a component of its production environment. Preferably, the antibody construct is free or substantially unrelated to all other components from its production environment. Contaminant components from its natural environment, such as those derived from recombinant transfected cells, are materials that may interfere with diagnostic or therapeutic applications of the polypeptide and may include enzymes, hormones, and other proteinaceous or non-proteinaceous solutes. The antibody constructs may comprise, for example, at least about 5% or at least about 50% of the total protein in a given sample. It is understood that the isolated protein may comprise from 5 to 99.9 wt.% of the total protein content, depending on the circumstances. The polypeptide can be produced at a significantly higher concentration by using an inducible promoter or a highly expressed promoter such that it is produced at higher concentrations. The definition includes the generation of an antibody construct in a wide variety of organisms and/or host cells that are known in the art. In preferred embodiments, the antibody construct will be purified (1) to a degree sufficient to obtain at least 15 residues on the N-terminal or internal amino acid sequence by using a rotating cup sequencer or (2) to homogeneity obtained by SDS-PAGE electrophoresis under non-reducing or reducing conditions using Coomassie blue or, preferably, silver staining. In this case, as a rule, the isolated antibody construct is obtained using at least one purification step.

The term binding domain characterizes in connection with the present invention a domain that (in particular) binds to/interacts with/recognizes a given target epitope or a given target site on CDH3 and CD3 target molecules (antigens), respectively. The structure and function of the first binding domain (recognizing CDH3) and preferably also the structure and/or function of the second binding domain (CD3) are/are based on the structure and/or function of an antibody, eg a full length or whole immunoglobulin molecule. According to the invention, the first binding

- 11 042775 domain is characterized by the presence of three light chain CDRs (ie CDR1, CDR2 and CDR3 of the VL region) and three heavy chain CDRs (ie CDR1, CDR2 and CDR3 of the VH region). The second binding domain preferably also contains the minimum structural requirements for the antibody, which allow binding to the target. More preferably, the second binding domain contains at least three light chain CDRs (ie CDR1, CDR2 and CDR3 of the VL region) and/or three CDRs of the heavy chain (ie CDR1, CDR2 and CDR3 of the VH region). It is contemplated that the first and/or second binding domain is or is generated by phage display or library screening methods and not by substitution of CDR sequences from a pre-existing (monoclonal) antibody into a scaffold.

In accordance with the present invention, the binding domains are preferably in the form of polypeptides. Such polypeptides may include proteinaceous portions and non-proteinaceous portions (eg, chemical linkers or chemical crosslinkers such as glutaraldehyde). Proteins (including their fragments, preferably biologically active fragments and peptides, usually having less than 30 amino acids) contain two or more amino acids linked to each other through a covalent peptide bond (to form an amino acid chain). As used herein, the term polypeptide describes a group of molecules that typically consist of more than 30 amino acids. The polypeptides may further form multimers such as dimers, trimers and higher oligomers, ie consisting of more than one polypeptide molecule. The polypeptide molecules forming such dimers, trimers, etc. may be identical or non-identical. The corresponding higher order structures of such multimers are therefore called homo- or heterodimers, homo- or heterotrimers, etc. An example for this heterocyclim is an antibody molecule, which in its natural form consists of two identical light polypeptide chains and two identical heavy polypeptide chains. The terms peptide, polypeptide and protein also refer to naturally modified peptides/polypeptides/proteins in which the modification is carried out, for example, by post-translational modifications such as glycosylation, acetylation, phosphorylation and the like. The peptide, polypeptide or protein referred to herein may also be chemically modified, such as pegylated. Such modifications are well known in the art and are described below.

As mentioned above, the binding domain may typically contain an antibody light chain variable region (VL) and an antibody heavy chain variable region (VH), however it need not contain both. Fd fragments, for example, have two VH regions and often retain some of the antigen-binding function of an intact antigen-binding domain. Examples of (modified) antigen-binding antibody fragments include (1) Fab fragment, a monovalent fragment having VL, VH, CL and CH1 domains; (2) an F(ab')2 fragment, a divalent fragment having two Fab fragments connected by a disulfide bridge at the hinge region; (3) an Fd fragment having two VH and CH1 domains; (4) an Fv fragment having VL and VH domains of one antibody arm, (5) a dAb fragment (Ward et al., (1989) Nature 341: 544-546) which has a VH domain; (6) isolated complementarity determination region (CDR); and (7) single chain Fv (scFv), the latter being preferred (eg, derived from an scFV library). Examples of embodiments of antibody constructs according to the invention are, for example, those described in WO 00/006605, WO 2005/040220, WO 2008/119567, WO 2010/037838, WO 2013/026837, WO 2013/026833, US 2014/0308285, US 2014/0302037, WO 2014/144722, WO 2014/151910 and WO 2015/048272.

Preferably the binding domain that binds to CDH3 and/or the binding domain that binds to CD3 is/are human binding domains. Antibodies and antibody constructs containing at least one human binding domain avoid some of the problems associated with antibodies or antibody constructs that have non-human variable and/or constant regions, such as rodent (e.g., mouse, rat , hamster or rabbit). The presence of such rodent-derived proteins may result in rapid clearance of the antibody or antibody construct, or may result in an immune response against the antibody or antibody construct by the patient. To avoid the use of rodent-derived antibodies or antibody constructs, fully human antibodies or antibody constructs can be made by introducing a human antibody function into the rodent such that the rodent produces fully human antibodies.

The ability to clone and regenerate human loci thousands of bases (megabases) in YAC and introduce them into the mouse germline provides a powerful approach to elucidating the functional components of very large or roughly mapped loci, as well as generating useful human disease models. In addition, the use of such technology to replace mouse loci with their human equivalents could provide unique insight into the expression and regulation of human gene products during development, their relationship to other systems, and their involvement in disease induction and progression.

An important practical application of such a strategy is the humanization of the mouse humoral immune system. The introduction of human immunoglobulin (Ig) loci into mice that have inactivated endogenous Ig genes provides an opportunity to study the mechanisms underlying programmed

- 12 042775 expression and assembly of antibodies, as well as their role in the development of B-cells. Moreover, such a strategy could be an ideal source for the production of fully human monoclonal antibodies (mAbs) - an important milestone in achieving promising antibody therapy in human diseases. Fully human antibodies or antibody constructs are expected to minimize the immunogenic and allergic reactions inherent in murine or mouse-derived monoclonal antibodies and thus improve the efficacy and safety of administered antibodies/antibody constructs. The use of fully human antibodies or antibody constructs can be expected to be a significant advantage in the treatment of chronic and recurrent human diseases such as inflammation, autoimmune diseases and cancer, which require repeated complex administrations.

One approach to this goal has been to construct mouse strains deficient in the production of mouse antibodies with large fragments of human Ig loci in the expectation that such mice will produce a large repertoire of human antibodies in the absence of mouse antibodies. Large fragments of human Ig retain a wide variety of variable genes, as well as the correct regulation of antibody production and expression. By using a murine mechanism to diversify and select antibodies and eliminate immunological tolerance to human proteins, the reproducible human antibody repertoire in these mouse strains should produce antibodies with high affinity for any antigen of interest, including human antigens. Using hybridoma technology, antigen-specific human mAbs with the desired specificity can be easily generated and selected. This general strategy has been demonstrated in connection with the production of the first strains of XenoMouse mice (see Green et al. Nature Genetics 7:13-21 (1994)). XenoMouse strains were constructed using yeast artificial chromosomes (YACs) containing 245 kb and 190 kb germline configuration fragments at the human heavy chain locus and the kappa light chain locus, respectively, which contained variable and constant region sequences. Human Ig containing YAC was found to be compatible with the mouse system for both rearrangement and antibody expression and was able to replace inactivated mouse Ig genes. This has been demonstrated by their ability to induce B cell development, produce an adult-like human repertoire of fully human antibodies, and generate antigen-specific human mAbs. These results also suggested that the introduction of large portions of human Ig loci containing more V genes, additional regulatory elements and human Ig constant regions could largely reproduce the full repertoire characteristic of the human humoral response to infection and immunization. The work of Green et al. has recently been expanded to include more than about 80% of the human antibody repertoire by introducing megabase size YAC fragments with the germline configuration of human heavy chain loci and kappa light chain loci, respectively. See Mendez et al. Nature Genetics 15:146-156 (1997) and US Patent Application No. 08/759620.

The production of XenoMouse mice is further discussed and characterized in US Patent Applications Nos. /464584, 08/464582, 08/463191, 08/462837, 08/486853, 08/486857, 08/486859, 08/462513, 08/724752 and 08/759620 and U.S. Patent Nos. 6,162,963; 6150584; 6114598; 6075181 and 5939598 and Japanese Patent Nos. 3068180 B2, 3068506 B2 and 3068507 B2. See also Mendez et al. Nature Genetics 15:146-156 (1997) and Green and Jakobovits J. Exp. Med. 188:483-495 (1998), EP 0463151 B1, WO 94/02602, WO 96/34096, WO 98/24893, WO 00/76310 and WO 03/47336.

In an alternative approach, others, including GenPharm International, Inc., have used the minilocus approach. In the mini-locus approach, an exogenous Ig locus is mimicked by incorporating pieces (individual genes) from the Ig locus. Thus, one or more VH genes, one or more DH genes, one or more JH genes, a mu constant region, and a second constant region (preferably a gamma constant region) are combined into a construct for administration to an animal. This approach is described in US Pat. No. 5,545,807 to Surani et al. and US Pat. Nos. 5,545,806; 5625825; 5625126; 5633425; 5661016; 5770429; 5789650; 5814318; 5877397; 5,874,299 and 6,255,458 each Lonberg and Kay US Pat. Nos. 5,591,669 and 6,023,010 Krimpenfort and Berns US Pat. 5721367 and 5789215 Berns et al., and U.S. Patent No. 5643763 Choi and Dunn, and GenPharm International U.S. Patent Application Nos. 08/053131, 08/096762, 08/155301, 08/161739, 08/165699, 08/209741. See also EP 0546073 B1, WO 92/03918, WO 92/22645, WO 92/22647, WO 92/22670, WO 93/12227, WO 94/00569, WO 94/25585, WO 96/14436, WO 97/ 13852 and WO 98/24884 and US Pat. No. 5,981,175. See also Taylor et al. (1992), Chen et al. (1993), Tuaillon et al. (1993), Choi et al. (1993), Lonberg et al. (1994), Taylor et al. (1994) and Tuaillon et al. (1995), Fishwild et al. (1996).

Kirin has also demonstrated the formation of human antibodies in mice in which large fragments of chromosomes or entire chromosomes have been introduced through microcell fusion. See European Patent Applications Nos. 773288 and 843961. Xenerex Biosciences is developing technology for the potential generation of human antibodies. In this technology, SCID mice are reconstituted with human lymphatic cells, such as B and/or T cells. Mice are then immunized with the antigen and can elicit an immune response against the antigen. See US patent No. 5476996; 5698767 and 5958765.

- 13 042775

Human anti-mouse antibody (HAMA) reactions have led the industry to prepare chimeric or otherwise humanized antibodies. However, it is expected that some human responses against the chimeric antibody (HACA) will be observed, especially when using chronic or multi-dose use of the antibody. Thus, it would be desirable to provide antibody constructs containing a fully human anti-CDH3 binding domain and a fully human anti-CD3 binding domain to enhance the effects and/or responses of HAMA or HACA.

The terms (specifically) binds to, (specifically) recognizes, (specifically) targets, and (specifically) reacts with means in accordance with the present invention that the binding domain interacts or specifically interacts with one or more, preferably at least two, more preferably at least three and most preferably at least four amino acids of the epitope located on the target protein or antigen (CDH3/CD3). The term epitope refers to a site on an antigen to which a binding domain, such as an antibody or immunoglobulin, or a derivative or fragment of an antibody or immunoglobulin, specifically binds. An epitope is antigenic, and therefore the term epitope is sometimes also referred to herein as an antigenic structure or antigenic determinant. Thus, the binding domain is the interaction site for the antigen. Said binding/interaction is also understood as specific recognition.

Epitopes can be formed by both adjacent amino acids and non-adjacent amino acids that converge as a result of protein folding into a tertiary structure. A linear epitope is an epitope in which the primary amino acid sequence contains a recognized epitope. A linear epitope usually contains at least 3 or at least 4 and more usually at least 5 or at least 6 or at least 7, for example 8 to 10 amino acids in a unique sequence, and may also be longer and contain at least 15 or 20 amino acids, at least 25 or 30 amino acids, or even more.

A conformational epitope, as opposed to a linear epitope, is an epitope in which the primary amino acid sequence containing the epitope is not the only defining component of the recognized epitope (eg, an epitope in which the primary amino acid sequence is not necessarily recognized by the binding domain). Typically, the conformational epitope contains an increased number of amino acids relative to the linear epitope. With regard to the recognition of conformational epitopes, the binding domain recognizes the three-dimensional structure of an antigen, preferably a peptide or protein, or fragment thereof (in the context of the present invention, the antigen for one of the binding domains is within CDH3). For example, when a protein molecule is folded to form a three-dimensional structure, certain amino acids and/or the polypeptide backbone that form the conformational epitope become matched, allowing the antibody to recognize the epitope. Methods for determining the conformation of epitopes include, but are not limited to, X-ray diffraction crystallography, two-dimensional nuclear magnetic resonance spectroscopy (2D-NMR), and site-directed spin labeling and electron paramagnetic resonance (EPR) spectroscopy.

The examples provided describe yet another method for characterizing a given binding domain, which includes checking whether a given binding domain binds to one or more epitope(s) of a given protein, in particular CDH3.

As used herein, the term epitope cluster refers to epitopes lying in a specific contiguous region of an antigen. An epitope cluster may contain one, two or more epitopes. The antibody construct may also bind to an epitope within the epitope cluster and in addition to another epitope outside the cluster, which may then correspond to a discontinuous epitope. For example, an antibody construct may bind to the extracellular subdomains of D3A and D3C, but not to D3B. The concept of epitope clustering is also used in characterizing the design features of the antibodies of the invention. The epitope cassettes and epitopes that have been defined in the context of the present invention in the extracellular domain of CDH3 are described above and shown in FIG. 1.

When the extracellular domain (D1-D5) or subdomain (A, B, C) thereof in a human CDH3 protein is replaced with the corresponding extracellular domain (D1-D5) or subdomain (A, B, C) of its human (e.g., chicken or mouse) CDH3 antigen ) (resulting in a construct containing human CDH3 in which one human extracellular domain or subdomain thereof is replaced by another non-human extracellular domain or subdomain thereof), there will be a decrease in binding domain binding. Said reduction is preferably at least 10%, 20%, 30%, 40%, 50%, more preferably at least 60%, 70%, 80%, 90%, 95% or even 100% compared to the corresponding epitope cassette in the human CDH3 protein, binding to the corresponding extracellular domain (D1-D5) or subdomain thereof in the human CDH3 protein is set to 100%. The aforementioned CDH3/non-human CDH3 chimeras are expected to be expressed in cells

- 14 042775

CHO. The chimeric human CDH3/non-human CDH3 is also expected to be fused to the transmembrane domain and/or the cytoplasmic domain of another membrane-associated protein, such as EpCAM.

A way to test for this loss of binding is by exchange with the appropriate extracellular domain (D1-D5) or subdomain thereof of a non-human (e.g. mouse, but others can also be assumed, such as rat, hamster, rabbit, chicken, etc.) CDH3- of the antigen is described in Example 2. Another way to determine the contribution of a particular target antigen residue to recognition by an antibody construct or binding domain is an alanine scan (see, e.g., Morrison KL & Weiss GA. Cur Opin Chem Biol. 2001 Jun; 5( 3):302-7), in which each residue to be analyzed is replaced by alanine, i.e. through site-directed mutagenesis. Alanya is used because of its small volume, chemically inert, functional group, which nevertheless mimics the secondary structure traits that many other amino acids have. Sometimes large amino acids such as valine or leucine can be used in cases where it is desirable to maintain the size of the mutated residues. Alanine scanning is not a new technology that has been used for a long period of time.

An interaction between a binding domain and an epitope or epitope cassette implies that the binding domain exhibits a marked affinity for the epitope or epitope cassette on a particular protein or antigen (in this case: CDH3 and CD3, respectively) and is generally not significantly reactive with proteins or antigens other than CDH3 or CD3. Marked affinity includes binding with an affinity of about 10 ^-6 M (KD) or greater. Preferably the binding is considered specific when the binding affinity is 10 ^-12 to 10 ^-8 M, 10 ^-12 to 10 ^-9 M, 10 ^-12 to 10 ^-10 M, 10 ^-11 to 10 ^-8 M, preferably from 10 ^-11 to 10 ^-9 M. Regardless of whether the binding domain specifically reacts with the target or binds to it, one can, in particular, compare the reaction of the specified binding domain with the target protein or antigen with the reaction of the specified binding domain with proteins or antigens other than CDH3 or CD3. Preferably, the binding domain of the invention does not substantially or substantially bind to proteins or antigens other than CDH3 or CD3 (i.e., the first binding domain is unable to bind to proteins other than CDH3 and the second binding domain is unable to bind to proteins, other than CD3).

The term does not substantially bind or is unable to bind means that the binding domain of the invention does not bind a protein or antigen other than CDH3 or CD3, i.e. does not show a reactivity of more than 30%, preferably not more than 20%, more preferably not more than 10% , particularly preferably not more than 9%, 8%, 7%, 6% or 5% with proteins or antigens other than CDH3 or CD3, with binding to CDH3 or CD3 respectively set to 100%.

Specific binding is believed to be mediated by specific motifs in the amino acid sequence of the binding domain and antigen. Thus, binding is achieved as a result of their primary, secondary and/or tertiary structure, as well as secondary modifications of these structures. The specific interaction of an interaction site of an antigen with its specific antigen may result in a simple binding of said site to the antigen. Moreover, the specific interaction of an antigen's interaction site with its specific antigen may alternatively or additionally lead to signal initiation, e.g., due to the induction of a conformational change in the antigen, oligomerization of the antigen, etc.

In another aspect, the invention provides a bispecific antibody construct comprising a first (preferably human) binding domain that binds to human CDH3 on the surface of a target cell and a second (preferably human) binding domain that binds to human CD3 on the surface of a T cell, wherein the first binding domain comprises a VH region containing CDR-H1, CDR-H2 and CDR-H3 and a VL region containing CDR-L1, CDRL2 and CDR-L3 selected from the group consisting of:

a) CDR-H1 as shown in SEQ ID NO: 149, CDR-H2 as shown in SEQ ID NO: 150, CDR-H3 as shown in SEQ ID NO: 151, CDR-L1 as shown in SEQ ID NO: 152, CDR-L2 as shown in SEQ ID NO: 153 and CDR-L3 as shown in SEQ ID NO: 154;

b) CDR-H1 as shown in SEQ ID NO: 159, CDR-H2 as shown in SEQ ID NO: 160, CDR-H3 as shown in SEQ ID NO: 161, CDR-L1 as shown in SEQ ID NO: 162, CDR-L2 as shown in SEQ ID NO: 163 and CDR-L3 as shown in SEQ ID NO: 164;

c) CDR-H1 as shown in SEQ ID NO: 169, CDR-H2 as shown in SEQ ID NO: 170, CDR-H3 as shown in SEQ ID NO: 171, CDR-L1 as shown in SEQ ID NO: 172, CDR-L2 as shown in SEQ ID NO: 173 and CDR-L3 as shown in SEQ ID NO: 174;

d) CDR-H1 as shown in SEQ ID NO: 179, CDR-H2 as shown in SEQ ID NO: 180, CDR-H3 as shown in SEQ ID NO: 181, CDR-L1 as shown in SEQ ID NO: 182, CDR-L2 as shown in SEQ ID NO: 183 and CDR-L3 as shown in SEQ ID NO: 184;

e) CDR-H1 as shown in SEQ ID NO: 189, CDR-H2 as shown in SEQ ID NO: 190, CDR-H3 as

- 15 042775 shown in SEQ ID NO: 191, CDR-L1 as shown in SEQ ID NO: 192, CDR-L2 as shown in SEQ ID NO: 193, and CDR-L3 as shown in SEQ ID NO: 194 ;

f) CDR-H1 as shown in SEQ ID NO: 199, CDR-H2 as shown in SEQ ID NO: 200, CDR-H3 as shown in SEQ ID NO: 201, CDR-L1 as shown in SEQ ID NO: 202, CDR-L2 as shown in SEQ ID NO: 203 and CDR-L3 as shown in SEQ ID NO: 204;

g) CDR-H1 as shown in SEQ ID NO: 209, CDR-H2 as shown in SEQ ID NO: 210, CDR-H3 as shown in SEQ ID NO: 211, CDR-L1 as shown in SEQ ID NO: 212, CDR-L2 as shown in SEQ ID NO: 213 and CDR-L3 as shown in SEQ ID NO: 214;

h) CDR-H1 as shown in SEQ ID NO: 219, CDR-H2 as shown in SEQ ID NO: 220, CDR-H3 as shown in SEQ ID NO: 221, CDR-L1 as shown in SEQ ID NO: 222, CDR-L2 as shown in SEQ ID NO: 223 and CDR-L3 as shown in SEQ ID NO: 224;

i) CDR-H1 as shown in SEQ ID NO: 229, CDR-H2 as shown in SEQ ID NO: 230, CDR-H3 as shown in SEQ ID NO: 231, CDR-L1 as shown in SEQ ID NO: 232, CDR-L2 as shown in SEQ ID NO: 233 and CDR-L3 as shown in SEQ ID NO: 234; And

j) CDR-H1 as shown in SEQ ID NO: 239, CDR-H2 as shown in SEQ ID NO: 240, CDR-H3 as shown in SEQ ID NO: 241, CDR-L1 as shown in SEQ ID NO: 242, CDR-L2 as shown in SEQ ID NO: 243 and CDR-L3 as shown in SEQ ID NO: 244.

In another aspect, the invention provides a bispecific antibody construct comprising a first (preferably human) binding domain that binds to human CDH3 on the surface of a target cell and a second (preferably human) binding domain that binds to human CD3 on the surface of a T cell, wherein the first binding domain comprises a VH region containing CDR-H1, CDR-H2 and CDR-H3 and a VL region containing CDR-L1, CDRL2 and CDR-L3 selected from the group consisting of:

a) CDR-H1 as shown in SEQ ID NO: 279, CDR-H2 as shown in SEQ ID NO: 280, CDR-H3 as shown in SEQ ID NO: 281, CDR-L1 as shown in SEQ ID NO: 282, CDR-L2 as shown in SEQ ID NO: 283 and CDR-L3 as shown in SEQ ID NO: 284;

b) CDR-H1 as shown in SEQ ID NO: 289, CDR-H2 as shown in SEQ ID NO: 290, CDR-H3 as shown in SEQ ID NO: 291, CDR-L1 as shown in SEQ ID NO: 292, CDR-L2 as shown in SEQ ID NO: 293 and CDR-L3 as shown in SEQ ID NO: 294;

c) CDR-H1 as shown in SEQ ID NO: 299, CDR-H2 as shown in SEQ ID NO: 300, CDR-H3 as shown in SEQ ID NO: 301, CDR-L1 as shown in SEQ ID NO: 302, CDR-L2 as shown in SEQ ID NO: 303 and CDR-L3 as shown in SEQ ID NO: 304;

d) CDR-H1 as shown in SEQ ID NO: 309, CDR-H2 as shown in SEQ ID NO: 310, CDR-H3 as shown in SEQ ID NO: 311, CDR-L1 as shown in SEQ ID NO: 312, CDR-L2 as shown in SEQ ID NO: 313 and CDR-L3 as shown in SEQ ID NO: 314;

e) CDR-H1 as shown in SEQ ID NO: 319, CDR-H2 as shown in SEQ ID NO: 320, CDR-H3 as shown in SEQ ID NO: 321, CDR-L1 as shown in SEQ ID NO: 322, CDR-L2 as shown in SEQ ID NO: 323 and CDR-L3 as shown in SEQ ID NO: 324;

f) CDR-H1 as shown in SEQ ID NO: 329, CDR-H2 as shown in SEQ ID NO: 330, CDR-H3 as shown in SEQ ID NO: 331, CDR-L1 as shown in SEQ ID NO: 332, CDR-L2 as shown in SEQ ID NO: 333 and CDR-L3 as shown in SEQ ID NO: 334;

g) CDR-H1 as shown in SEQ ID NO: 339, CDR-H2 as shown in SEQ ID NO: 340, CDR-H3 as shown in SEQ ID NO: 341, CDR-L1 as shown in SEQ ID NO: 342, CDR-L2 as shown in SEQ ID NO: 343 and CDR-L3 as shown in SEQ ID NO: 344; And

h) CDR-H1 as shown in SEQ ID NO: 349, CDR-H2 as shown in SEQ ID NO: 350, CDR-H3 as shown in SEQ ID NO: 351, CDR-L1 as shown in SEQ ID NO: 352, CDR-L2 as shown in SEQ ID NO: 353 and CDR-L3 as shown in SEQ ID NO: 354.

The term variable refers to portions of antibodies or immunoglobulin domains that exhibit variability in their sequence and that are involved in determining the specificity and binding affinity of a particular antibody (ie, variable domain(s)). The pairing of the variable heavy chain (VH) and the variable light chain (VL) together form one antigen-binding site. The CH domain closest to VH is designated as CH1. Each L chain is linked to an H chain by one covalent disulfide bond, while two H chains are linked to each other by one or more disulfide bonds, depending on the isotype of the H chain.

Variability is distributed unevenly across all variable domains of antibodies; it is concentrated in subdomains of each of the heavy and light chain variable regions. These subdomains are called hypervariable regions or complementarity determining regions (CDRs). The more conserved (ie, non-hypervariable) portions of the variable domains are called framework regions (FRMs or FRs) and provide the basis for six CDRs in three-dimensional space to form an antigen-binding surface. Each of the native heavy and light chain variable domains contains four FRM regions (FR1, FR2, FR3, and FR4) predominantly adopting a beta-sheet configuration, connected by three hypervariable regions that form

- 16 042775 loops connecting the structure of the beta sheet, and in some cases forming part of the structure of the beta sheet. The hypervariable regions of each chain are joined to each other in close proximity by FRM and, together with the hypervariable regions of the other chain, participate in the formation of an antigen-binding site (see Kabat et al., loc. cit). The constant domains are not directly involved in antigen binding, but exhibit various effector functions such as, for example, antibody-dependent, cell-mediated cytotoxicity and complement activation.

The terms CDR and its multiple CDRs refer to the domain of complementarity, of which three make up the binding character of the light chain variable region (CDR-L1, CDR-L2 and CDR-L3) and three make up the binding character of the heavy chain variable region (CDR-H1, CDR -H2 and CDR-H3). CDRs contain most of the residues responsible for specific antibody-antigen interactions and therefore contribute to the functional activity of the antibody molecule: they are the main determinants of antigen specificity.

The specific boundaries and lengths of CDRs are defined by different classification and numbering systems. Therefore, CDRs may refer to Kabat, Chothia, contact, or any other boundary definition, including the numbering system described in this document. Despite different boundaries, each of these systems has some degree of overlap in what constitutes the so-called hypervariable regions in variable sequences. Thus, CDR definitions according to these systems may differ in length and boundary regions relative to the adjacent frame region. See, for example, Kabat (cross-species sequence variability approach), Chothia (antigen-antibody complex crystallographic approach), and/or MacCallum (Kabat et al., loc. cit.; Chothia et al., J Mol. Biol, 1987, 196:901-917 and MacCallum et al., J. Mol. Biol, 1996, 262:732). Another standard for characterizing an antigen binding site is the AbM definition used by Oxford Molecular's AbM antibody modeling software. See, for example, Protein Sequence and Structure Analysis of Antibody Variable Domains. E: Antibody Engineering Lab Manual (Ed.: Duebel, S. and Kontermann, R., Springer-Verlag, Heidelberg). To the extent that the two residue identification methods define regions of overlapping but not identical regions, they can be combined to determine a hybrid CDR. However, numbering according to the so-called Kabat system is preferred.

As a rule, CDRs form a loop structure, which can be classified as a canonical structure. The term canonical structure refers to the backbone conformation that is accepted by antigen binding loops (CDRs). From comparative structural studies, five of the six antigen-binding loops were found to have only a limited repertoire of available conformations. Each canonical structure can be characterized by the twist angles of the polypeptide backbone. Thus, correlation loops between antibodies can have very similar three-dimensional structures despite high amino acid sequence variability in most parts of the loops (Chothia and Lesk, J. Mol. Biol, 1987, 196: 901; Chothia et al., Nature, 1989, 342: 877; Martin and Thornton, J. Mol. Biol, 1996, 263: 800). In addition, there is a relationship between the adopted loop structure and its amino acid sequences. The conformation of a particular canonical class is determined by the length of the loop and the amino acid residues found at key positions within the loop as well as within the conserved structure (i.e., outside the loop). Thus, the assignment of a particular canonical class can be based on the presence of these key amino acid residues.

The term canonical structure may also include considerations regarding the linear sequence of an antibody, for example, as cataloged in Kabat (Kabat et al., loc. lit.). The Kabat numbering scheme is a widely accepted standard for numbering the amino acid residues of an antibody variable domain in a consistent manner and is the preferred scheme used in this invention, as also mentioned elsewhere herein. Additional structural considerations may also be used to determine the canonical structure of an antibody. For example, these differences, not fully reflected by Kabat's numbering, can be described by the numbering system of Chothia et al. and/or identified by other methods, such as crystallography and two- or three-dimensional computational modeling. Accordingly, a given antibody sequence can be placed in a canonical class that allows, among other things, the identification of the corresponding chassis sequences (eg, based on the desire to include multiple canonical structures in the library). Kabat numbering of antibody amino acid sequences and structural considerations as described by Chothia et al., loc. cit. and their implications for the interpretation of canonical aspects of antibody structure are described in the literature. The subunit structure and three-dimensional configuration of various classes of immunoglobulins are well known in the art. For a review of the structure of antibodies, see Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory, eds. Harlow et al., 1988.

The light chain CDR3, and in particular the heavy chain CDR3, can form the most important antigen-binding determinants within the light and heavy chain variable regions. In some antibody constructs, the heavy chain CDR3 constitutes the main contact area between the antigen

- 17 042775 number and antibody. In vitro selection schemes that vary only CDR3 can be used to vary the binding properties of an antibody or to determine which residues contribute to antigen binding. Therefore, CDR3 is typically the largest source of molecular diversity at an antibody binding site. H3, for example, can be shorter than two amino acid residues or have more than 26 amino acids.

The sequence of antibody genes after assembly and somatic mutation is highly variable, and these diverse genes are estimated to encode 10 different antibody molecules (Immunoglobulin Genes, ^2nd ed., eds. Jonio et al., Academic Press, San Diego, CA, 1995) . Accordingly, the immune system provides a repertoire of immunoglobulins. The term repertoire refers to at least one nucleotide sequence derived in whole or in part from at least one sequence encoding at least one immunoglobulin. The sequence(s) can be generated by in vivo rearrangement of the V, D and J segments of the heavy chains and the V and J segments of the light chains. Alternatively, the sequence(s) can be obtained from a cell in response to which a rearrangement occurs, for example, stimulation in vitro. Alternatively, part or all of the sequence(s) may be obtained by DNA splicing, nucleotide synthesis, mutagenesis and other methods, see, for example, US patent No. 5565332. The repertoire may include only one sequence or may include multiple sequences, including genetically diverse collection.

In one embodiment, the first binding domain of an antibody construct of the invention comprises a VH region selected from the group consisting of VH regions as shown in SEQ ID NO: 155, SEQ ID NO: 165, SEQ ID NO: 175, SEQ ID NO: 185, SEQ ID NO: 195, SEQ ID NO: 205, SEQ ID NO: 215, SEQ ID NO: 225, SEQ ID NO: 235, and SEQ ID NO: 245.

In yet another embodiment of an antibody construct of the invention, the first binding domain comprises a VL region selected from the group consisting of VL regions as shown in SEQ ID NO: 156, SEQ ID NO: 166, SEQ ID NO: 176, SEQ ID NO: 186, SEQ ID NO: 196, SEQ ID NO: 206, SEQ ID NO: 216, SEQ ID NO: 226, SEQ ID NO: 236 and SEQ ID NO: 246.

In another embodiment of an antibody construct of the invention, the first binding domain comprises a VH region and a VL region selected from the group consisting of VH region and VL region pairs as shown in SEQ ID NO: 155+156, SEQ ID NO: 165+166, SEQ ID NO: 175+176, SEQ ID NO: 185+186, SEQ ID NO: 195+196, SEQ ID NO: 205+206, SEQ ID NO: 215+216, SEQ ID NO: 225+226, SEQ ID NO: 235+236 and SEQ ID NO: 245+246.

In another embodiment, the first binding domain of an antibody construct of the invention comprises a VH region selected from the group consisting of VH regions as shown in SEQ ID NO: 285, SEQ ID NO: 295, SEQ ID NO: 305, SEQ ID NO: 315, SEQ ID NO: 325, SEQ ID NO: 335, SEQ ID NO: 345, and SEQ ID NO: 355.

In yet another embodiment of an antibody construct of the invention, the first binding domain comprises a VL region selected from the group consisting of VL regions as shown in SEQ ID NO: 286, SEQ ID NO: 296, SEQ ID NO: 306, SEQ ID NO: 316, SEQ ID NO: 326, SEQ ID NO: 336, SEQ ID NO: 346 and SEQ ID NO: 356.

In another embodiment of an antibody construct of the invention, the first binding domain comprises a VH region and a VL region selected from the group consisting of VH region and VL region pairs as shown in SEQ ID NO: 285+286, SEQ ID NO: 295+296, SEQ ID NO: 305+306, SEQ ID NO: 315+316, SEQ ID NO: 325+326, SEQ ID NO: 335+336, SEQ ID NO: 345+346 and SEQ ID NO: 355+356.

As used herein, the term bispecific refers to an antibody construct that is at least bispecific, i.e. it contains at least a first binding domain and a second binding domain, the first binding domain binding to a single antigen or target (herein: CDH3) , and the second binding domain binds to another antigen or target (in this document: CD3). Accordingly, the antibody constructs of the invention contain specificities for at least two different antigens or targets. The term bispecific antibody construct of the invention also encompasses multispecific antibody constructs, such as trispecific antibody constructs, the latter of which include three binding domains or constructs having more than three (eg four, five...) specificities.

Given that the antibody constructs according to the invention are (at least) bispecific, they do not naturally occur and they differ markedly from natural products. Thus, a bispecific antibody construct or immunoglobulin is an artificial fusion antibody or immunoglobulin having at least two different binding sites with different specificities. Bespecifically antibodies can be obtained in various ways, including through the fusion of hybridomas or linking Fab' fragments. See, for example, Songsivilai & Lachmann, Clin. Exp. Immunol. 79:315-321 (1990).

The at least two binding domains and variable domains of an antibody construct of the invention may or may not contain peptidic linkers (spacer peptides). Ter

- 18 042775 min peptide linker defines in accordance with the present invention the amino acid sequence by which the amino acid sequences of one (variable and/or binding) domain and another (variable and/or binding) domain of the antibody construct of the invention are linked to each other. An essential technical feature of such a peptide linker is that it does not contain any polymerization activity. Suitable peptide linkers include those described in US Pat. Nos. 4,751,180 and 4,935,233 or WO 88/09344.

Where a linker is used, the linker is preferably of sufficient length and sequence to ensure that each of the first and second domains can independently maintain their differential binding specificities. For peptide linkers that connect at least two binding domains in an antibody construct of the invention (or two variable domains), those peptide linkers are preferred that contain only a few amino acid residues, eg, 12 amino acid residues or less. Thus, a peptide linker of 12, 11, 10, 9, 8, 7, 6 or 5 amino acid residues is preferred. A putative peptide linker with less than 5 amino acids contains 4, 3, 2, or one amino acid(s), with Gly-rich linkers being preferred. A particularly preferred single amino acid in the context of said peptide linker is Gly. Suitably, said peptide linker may consist of a single Gly amino acid. Another preferred embodiment of the peptide linker is characterized by the amino acid sequence Gly-Gly-Gly-Gly-Ser, ie Gly ₄ Ser (SEQ ID NO: 393), or polymers thereof, ie. (Gly ₄ Ser) _x , where x is an integer equal to 1 or more. The characteristics of said peptide linker, which include the lack of promotion of secondary structures, are known in the art and are described, for example, in Dall'Acqua et al. (Biochem. (1998) 37, 9266-9273), Cheadle et al. (Mol Immunol (1992) 29, 21-30) and Raag and Whitlow (FASEB (1995) 9(1), 73-80). Preferred are peptide linkers which also do not promote the formation of any secondary structures. Linking these domains to each other can be achieved, for example, using genetic engineering, as described in the examples. Methods for preparing fused and operably linked bispecific single strand constructs and expressing them in bacterial or mammalian cells are well known in the art (e.g., WO 99/54440 or Sambrook et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 2001).

Thus, the present invention provides a preferred embodiment wherein the antibody construct is in a format selected from the group consisting of (scFv)2, single domain-scFv mAb, diabody, and oligomers of any of the above formats. The term is in the format does not exclude that the design can be further modified, for example, by attaching or merging with other fragments, as described in this document.

In accordance with a particularly preferred embodiment of the invention, the antibody construct of the invention is a bispecific single chain antibody construct, more preferably a bispecific single chain Fv (scFv). Although the two domains of the Fv fragment, VL and VH, are encoded by separate genes, they can be combined using standard recombinant techniques through a synthetic linker, which allows them to be obtained as a single protein chain in which the VL and VH regions are paired with the formation of a monovalent molecule; see, for example, Huston et al. (1988) Proc. Natl. Acad. Sci USA 85:5879-5883). These antibody fragments are prepared using conventional methods known to those skilled in the art and the fragments are evaluated for function in the same manner as whole or full length antibodies.

A single chain variable fragment (scFv) is therefore an immunoglobulin heavy chain (VH) and light chain (VL) variable region fusion protein, typically associated with a short linker peptide of about 10 to about 25 amino acids, preferably about 15 to 20 amino acids. The linker is usually rich in glycine for flexibility and serine or threonine for solubility, and can either connect the N-terminus of the VH to the C-terminus of the VL, or vice versa. This protein retains the specificity of the original immunoglobulin despite the removal of constant regions and the introduction of a linker.

Bispecific single chain molecules are known in the art and are described in WO 99/54440, Mack, J. Immunol. (1997), 158, 3965-3970, Mack, PNAS, (1995), 92, 7021-7025, Kufer, Cancer Immunol. Immunother., (1997), 45, 193-197, Loffler, Blood, (2000), 95, 6, 2098-2103, Brhhl, Immunol., (2001), 166, 2420-2426, Kipriyanov, J. Mol. Biol., (1999), 293, 41-56. The methods described for the production of single chain antibodies (see, in particular, US patent No. 4946778, Kontermann and Dubel (2010), loc. cit. and Little (2009), loc. cit) can be adapted to obtain single chain antibody constructs specifically recognizing the chosen target(s).

Bivalent (also called bivalent) or bispecific single chain variable fragments (bi-scFvs or di-scFvs, having the (scFv)2 format, can be constructed by linking two scFv molecules. If the two scFv molecules have the same binding specificity, the resulting (scFv) 2 molecule will preferably be called bivalent (i.e. it has two valences for the same target epitope.) If two scFv molecules have different

- 19 042775 binding specificity, the resulting (scFv)2 molecule will preferably be referred to as bispecific. Coupling can be accomplished by creating a single peptide chain with two VH regions and two VL regions, resulting in tandem scFvs (see, for example, Kufer P. et al., (2004) Trends in Biotechnology 22(5):238-244). Another possibility is to design scFv molecules with linker peptides that are too short for the two variable regions to add up (eg, about five amino acids), causing the scFvs to dimerize. This type is known as the diabody (see, for example, Hollinger, Philipp et al., (July 1993) Proceedings of the National Academy of Sciences of the United States of America 90 (14): 6444-8.).

In yet another preferred embodiment of an antibody of the invention, the heavy chain (VH) and light chain (VL) of the binding domain (binding to either the target CDH3 antigen or CD3) are not directly linked via a peptide linker as described above, but the binding domains formed as described for the diabody. Thus, a VH of a CD3 binding domain can be fused to a VL of a CDH3 binding domain using a peptide linker, and a VH of a CDH3 binding domain is fused to a VL of a CD3 binding domain via such a peptide linker.

Single domain antibodies contain only one (monomeric) antibody variable domain, which is capable of selectively binding to a particular antigen independently of other V regions or domains. The first single domain antibodies were constructed from heavy chain antibodies found in camelids and these are called V _H H fragments. Cartilaginous fish also have heavy chain antibodies (IgNAR) from which single domain antibodies called V _NAR fragments can be derived. An alternative approach is to separate the dimeric variable domains from common immunoglobulins, eg, from human or rodent, into monomers, hence obtaining VH or VL as a single domain Ab. While most research on single-domain antibodies to date is based on heavy chain variable domains, light chain-derived nanoparticles have been shown to specifically bind to targeted epitopes. Examples of single domain antibodies are called sdAbs, nanoparticles, or single variable domain antibodies.

Thus (single domain mAb) ₂ is a monoclonal antibody construct consisting of (at least) two single domain monoclonal antibodies individually selected from the group consisting of VH, VL, VHH and V _NAR . The linker is preferably in the form of a peptide linker. Similarly, a scFv single domain mAb is a monoclonal antibody consisting of at least one single domain antibody as described above and one scFv molecule as described above. Again, the linker is preferably in the form of a peptide linker.

In one embodiment, the first binding domain comprises an amino acid sequence selected from the group consisting of those sequences shown in SEQ ID NO: 157, SEQ ID NO: 167, SEQ ID NO: 177, SEQ ID NO: 187, SEQ ID NO: 197, SEQ ID NO: 207, SEQ ID NO: 217, SEQ ID NO: 227, SEQ ID NO: 237 and SEQ ID NO: 247.

In another embodiment, the first binding domain comprises an amino acid sequence selected from the group consisting of those sequences shown in SEQ ID NO: 287, SEQ ID NO: 297, SEQ ID NO: 307, SEQ ID NO: 317, SEQ ID NO: 327, SEQ ID NO: 337, SEQ ID NO: 347 and SEQ ID NO: 357.

It is also contemplated that the antibody construct of the invention has, in addition to its function of binding to CDH3 and CD3 target molecules, another function. In this format, an antibody construct is a trifunctional or multifunctional antibody construct that functions by targeting target cells via CDH3 binding, mediating T cell cytotoxic activity via CD3 binding, and providing an additional function such as antibody-dependent cellular cytotoxicity mediated by a fully functional an Fc constant domain, by recruiting effector cells such as NK cells, labels (fluorescent, etc.), therapeutic agents such as a toxin or radionuclide, and/or agents for increasing serum half-life, etc.

Examples of methods for extending the serum half-life of the antibody constructs of the invention include peptides, proteins, or protein domains that are fused or otherwise attached to the antibody constructs. The group of peptides, proteins or protein domains includes peptides that bind to other proteins with a preferred pharmacokinetic profile in the human body, such as serum albumin (see WO 2009/127691). One of these is shown in SEQ ID NO: 437. An alternative concept of such extended half-life (HLE) peptides includes peptides that bind to the neonatal Fc receptor (FcRn, see WO 2007/098420), which are also used in some of the constructs of this invention. . The concept of attaching larger domains of proteins or complete proteins includes, for example, the fusion of human serum albumin, variants or mutants of human serum albumin (see WO 2011/051489, WO 2012/059486, WO 2012/150319, WO 2013/135896, WO 2014/ 072481, WO 2013/075066) or their domains, as well as the fusion of the constant region of immunoglobulins (Fc domains) and their variants. Such Fc domain variants can be optimized/modified to provide the desired pairing of dimers or multimemes.

- 20 042775 ditch, to prevent binding of the Fc receptor (eg Fc receptor) or for other reasons. If the HLE molecules described above consist of only one single polypeptide chain, they have the advantage that (i) there is no need for two separate expression systems, and (ii) they can be isolated in high purity due to the absence of a dummy chain. Another concept known in the art for extending the half-life of small protein compounds in the human body is the pegylation of these compounds, such as the antibody construct of the present invention.

In a preferred embodiment, the bispecific antibody constructs of the invention may be linked (eg, via a peptide bond) to a fusion partner (such as a protein or polypeptide or peptide), for example, to extend the serum half-life of the construct. These fusion partners can be selected from human serum albumin (HSA or HALB) as its sequence variants, HSA binding peptides, FcRn binding peptides (FcRn BP), or constructs containing an (antibody derived) Fc region. Illustrative sequences of these fusion partners are shown in SEQ ID NOs: 406-421 and 437-444. In general, fusion partners can be linked to the N-terminus or C-terminus of the bispecific antibody constructs of the invention either directly (e.g. via a peptide bond) or via a peptide linker such as (GGGGS) _n (where n is an integer a number equal to 2 or more, such as 2 or 3 or 4). Suitable peptide linkers are shown in SEQ ID NO: 392-400.

An antibody construct named CDH3-13 (the full-length sequence of the bispecific molecule shown in SEQ ID NO: 178) was fused in frame to a group of fusion proteins or fusion peptides (see e.g. SEQ ID NO: 437-444) to extend the duration serum half-life of the construct. The respective sequences of these fusion constructs are shown in SEQ ID NOs: 379-389. The sequences of these naked bispecific antibody construct fusion partners can also be linked (C- or N-terminally, as appropriate and shown for CDH3-13) to any of the other antibody constructs described herein. Therefore, it is contemplated that the bispecific antibody construct of the invention further comprises a polypeptide as set forth in SEQ ID NO: 437 or albumin, preferably human albumin or a variant thereof (having improved properties such as affinity for the FcRn receptor and a long plasma half-life), most preferably albumin as shown in SEQ ID NO: 443 or 444. These fragments are preferably fused in frame to the C-terminus of the bispecific antibody construct.

Example 15 also shows the unexpected advantage of C-terminal fusion of albumin to the bispecific antibody construct of the invention. For these bispecific T cell interacting molecules containing a binding domain specific for human and Callithrix jacchus, Saguinus oedipus or Saimiri sciureus CD3ε, where the epitope is part of an amino acid sequence consisting of the group consisting of SEQ ID NOs: 2, 4, 6, or 8 WO 2008/119567 and contains at least the amino acid sequence Gln-Asp-Gly-Asn-Glu, these molecules, when used at very high concentrations, have been found to exhibit T-cell cytotoxicity even in the absence of target cells . Such high concentration issues may become relevant for specific routes of administration or in combination with specific target settings and required compound concentrations. Preferred examples of such second binding domains that bind to human CD3 on the surface of a T cell are described below and are shown in SEQ ID NOs: 445-537. When serum albumin is fused to the C-terminus of such a bispecific construct, T cell cytotoxicity is avoided, see FIG. 13. Without intending to be bound by theory, T cell activation in the presence of a high concentration of T cells interacting with bispecific antibody constructs and in the absence of target cells can be explained by dimerization or multimerization of antibody constructs via the CD3 binding domain. Such di- or multimerization is sterically disrupted by fusing albumin or a variant thereof to the C-terminus of the antibody construct, while retaining the characteristics of the antibody construct for the mode of activation of interaction with a T cell. Therefore, a preferred antibody construct of this invention contains in the N-C terminal order:

a first binding domain having an amino acid sequence selected from the group consisting of SEQ ID NO: 157, SEQ ID NO: 167, SEQ ID NO: 177, SEQ ID NO: 187, SEQ ID NO: 197, SEQ ID NO: 207, SEQ ID NO: 217, SEQ ID NO: 227, SEQ ID NO: 237, SEQ ID NO: 247, SEQ ID NO: 287, SEQ ID NO: 297, SEQ ID NO: 307, SEQ ID NO: 317, SEQ ID NO: 327, SEQ ID NO: 337, SEQ ID NO: 347 and SEQ ID NO: 357;

a peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NO: 392-400;

a second binding domain having an amino acid sequence selected from the group consisting of SEQ ID NO: 453, SEQ ID NO: 462, SEQ ID NO: 471, SEQ ID NO: 480, SEQ ID NO: 489, SEQ ID NO: 498, SEQ ID NO: 507, SEQ ID NO: 516, SEQ ID NO: 525, SEQ ID NO: 534 and SEQ ID NO: 537; And

- 21 042775 optionally, a peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NO: 392-400; and a polypeptide having an amino acid sequence selected from the group consisting of SEQ ID NOs: 437, 443 and 444.

In accordance with another preferred embodiment of the invention, the bispecific antibody construct of the invention contains (in addition to the two binding domains) a third domain that contains two polypeptide monomers, each of which contains a hinge, a CH2 and a CH3 domain, these two polypeptides (or polypeptides monomers) are fused to each other via a peptide linker. Preferably, said third domain comprises, in the N-C terminal order: hinge-CH2-CH3-linker-hinge-CH2-CH3. Preferred amino acid sequences for this third domain are shown in SEQ ID NOs: 414-421. Each of these polypeptide monomers preferably has an amino acid sequence that is selected from the group consisting of SEQ ID NO: 406-413, or at least 90% identical to those sequences. In another preferred embodiment, the first and second binding domains of the bispecific antibody construct of the invention are fused to the third domain via a peptide linker that is, for example, selected from the group consisting of any of SEQ ID NOs: 392-400, preferably from the group consisting of any of SEQ ID NOs: 392, 393, 395, 396, 397, 399 and 400.

In accordance with the present invention, the hinge is an IgG hinge region. This site can be identified by analogy using Kabat numbering, see Kabat entries 223-243. In accordance with the above minimum requirement for the hinge are amino acid residues corresponding to the plot of the IgG1 sequence from D231 to P243 in accordance with Kabat numbering. The terms CH2 and CH3 refer to constant regions 2 and 3 of an immunoglobulin heavy chain. These sites can also be identified by analogy using Kabat numbering, see Kabat entries 244-360 for CH2 and Kabat entries 361-478 for CH3. It is understood that there is some variation among immunoglobulins between their IgG1 Fc region, IgG2 Fc region, IgG3 Fc region, IgG4 Fc region, IgM Fc region, IgA Fc region, IgD Fc region, and IgE Fc region (see, e.g., Padlan, Molecular Immunology , 31(3), 169-217 (1993)). The term Fc monomer refers to the last two IgA heavy chain constant regions, IgD and IgG, and the last three IgE and IgM heavy chain constant regions. The Fc monomer may also contain a flexible N-terminal hinge to these domains. For IgA and IgM, the Fc monomer may contain a J chain. In IgG, the Fc portion contains the immunoglobulin domains CH2 and CH3 and a hinge between the first two domains and CH2. Although the boundaries of the Fc portion of an immunoglobulin may vary, an example for a human IgG heavy chain Fc portion containing a functional hinge, a CH2 domain and a CH3 domain can be defined, for example, containing residues D231 (hinge domain) to P476 (C-terminus of CH3 domain) or respectively from D231 to L476 for IgG4, where the numbering corresponds to Kabat.

An antibody construct of the invention may therefore contain, in order from N- to C-terminus:

(a) a first binding domain;

(b) a peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NOs: 393, 399 and 400;

(c) a second binding domain;

(d) a peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NOs: 392, 393, 395, 396, 397, 399 and 400;

(e) a first third domain polypeptide monomer (comprising a hinge, CH2 and a CH3 domain);

(f) a peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NOs: 402, 403, 404 and 405; and (g) a second third domain polypeptide monomer (comprising a hinge, CH2 and a CH3 domain);

It is also preferred that the antibody construct of the invention contains, in order from N to C-terminus:

(a) a first binding domain having an amino acid sequence selected from the group consisting of SEQ ID NO: 157, SEQ ID NO: 167, SEQ ID NO: 177, SEQ ID NO: 187, SEQ ID NO:

197, SEQ ID NO: 207, SEQ ID NO: 217, SEQ ID NO: 227, SEQ ID NO: 237, SEQ ID NO: 247, SEQ ID NO:

287, SEQ ID NO: 297, SEQ ID NO: 307, SEQ ID NO: 317, SEQ ID NO: 327, SEQ ID NO: 337, SEQ ID NO:

347 and SEQ ID NO: 357;

(b) a peptide linker having an amino acid sequence selected from the group consisting of SEQ ID NOs: 393, 399 and 400;

(c) a second binding domain having an amino acid sequence selected from the group consisting of SEQ ID NO: 453, SEQ ID NO: 462, SEQ ID NO: 471, SEQ ID NO: 480, SEQ ID NO: 489, SEQ ID NO : 498, SEQ ID NO: 507, SEQ ID NO: 516, SEQ ID NO: 525, SEQ ID NO: 534 and SEQ ID NO: 537;

(d) a peptide linker having an amino acid sequence selected from the group

- 22 042775 standing for SEQ ID NOs: 392, 393, 395, 396, 397, 399 and 400; and (e) a third domain having an amino acid sequence selected from the group consisting of SEQ ID NOs: 414-421.

Therefore, in a preferred embodiment, the antibody construct of the invention comprises or consists of a polypeptide selected from the group consisting of those shown in SEQ ID NO: 422, SEQ ID NO: 423, SEQ ID NO: 424, SEQ ID NO : 425, SEQ ID NO: 426, SEQ ID NO: 427, SEQ ID NO: 428, SEQ ID NO: 429, SEQ ID NO: 430, SEQ ID NO: 431, SEQ ID NO: 432, SEQ ID NO: 433 , SEQ ID NO: 434 and SEQ ID NO: 435.

Reference is also made to Example 15, which shows one of the above Fc constructs and its antitumor activity in a mouse xenograft model. Covalent modifications of antibody constructs are also included within the scope of this invention and are usually, but not always, performed post-translationally. For example, several types of covalent modifications to the antibody construct are introduced into the molecule by reacting specific amino acid residues of the antibody construct with an organic derivatizing agent that is capable of reacting with selected N- or C-terminal side chains.

Cysteinyl residues are most commonly reacted with α-haloacetates (and the corresponding amines) such as chloroacetic acid or chloroacetamide to give carboxymethyl or carboxyamidomethyl derivatives. Cysteinyl residues are also derivatized by reaction with bromotrifluoroacetone, α-bromo-β-(5-imidozoyl)propionic acid, chloroacetyl phosphate, N-alkylmaleimides, 3-nitro-2-pyridyl disulfide, methyl-2-pyridyl disulfide, p-chloromercury benzoate, 2-chloromercury -4-nitrophenol or chloro-7-nitrobenzo-2-oxa-1,3-diazole.

Histidyl residues are derivatized by reaction with diethyl pyrocarbonate at pH 5.5-7.0 because this agent is relatively specific for the histidyl side chain. Para-bromophenacyl bromide is also applicable; the reaction is preferably carried out in 0.1 M sodium cacodylate solution at pH 6.0. Lysinyl and amino-terminal residues are reacted with succinic anhydride or anhydrides of other carboxylic acids. Derivatization with these agents can lead to a change in the charge of lysinyl residues. Other suitable reagents for derivatizing alpha-amino residues include imidoesters such as methyl picolinimidate; pyridoxal phosphate; pyridoxal; chloroborohydride; trinitrobenzenesulfonic acid; O-methylisourea; 2,4pentanedione; and a transaminase-catalyzed reaction with glyoxylate.

Arginyl residues are modified by reaction with one or more conventional reagents, including phenylglyoxal, 2,3-butanedione, 1,2-cyclohexanedione, and ninhydrin. Derivatization of arginine residues requires that the reaction be carried out in an alkaline environment due to the high pKa of the guanidine functional group. In addition, such reagents can interact with lysine groups, as well as with the arginine epsilon-amino group.

Specific modifications of tyrosyl residues can be carried out, of particular interest are the cases of introducing spectral labels into tyrosyl residues by reaction with aromatic diazonium compounds or tetranitromethane. The most commonly used are N-acetylimidizole and tetranitromethane to form O-acetyltyrosyl and 3-nitro derivatives, respectively. Tyrosyl residues are iodinated using ¹²⁵ I or ¹³¹ I to produce labeled proteins for use in radioimmunoassay, the chloramine T method described above being suitable.

Carboxyl side groups (aspartyl or glutamyl) are selectively modified by reaction with carbodiimides (R'-N=C=N--R'), where R and R' are optionally different alkyl groups such as 1-cyclohexyl-3- (2-morpholinyl-4-ethyl)carbodiimide or 1-ethyl-3-(4-azonia4,4-dimethylphenyl)carbodiimide. In addition, aspartyl and glutamyl residues are converted to asparaginyl and glutaminyl residues by reaction with ammonium ions.

Derivatization with bifunctional agents is used to cross-link antibody constructs of this invention with a water-insoluble matrix substrate or surface for use in various methods. Commonly used cross-linking agents include, for example, 1,1-bis(diazoacetyl)-2-phenylethane, glutaraldehyde, N-hydroxysuccinimide esters, for example, esters with 4-azidosalicylic acid, homobifunctional imidoesters, including disuccinimidyl esters, such as 3,3 '-dithio-bis-(succinimidyl propionate), and bifunctional maleimides such as bis-M-maleimido-1,8-octane. Using derivatizing agents such as methyl 3-[(p-azidophenyl)dithio]propioimidate, photoactivated intermediates are obtained which are capable of forming crosslinks in the presence of light. Alternatively, water-insoluble matrices such as cyanogen bromide-activated carbohydrates and reactive substrates are used to immobilize the protein, as described in US Pat. Nos. 3,969,287; 3691016; 4195128; 4247642; 4229537 and 4330440.

The glutaminyl and asparaginyl residues are often deamidated to the corresponding glutamyl and aspartyl residues, respectively. Alternatively, these residues are deamidated under moderately acidic conditions. Any form of these residues is within the scope of this invention.

- 23 042775

Other modifications include hydroxylation of proline and lysine, phosphorylation of hydroxyl groups of seryl or threonyl residues, methylation of a-amino groups of lysine, arginine, and histidine side chains (T. E. Creighton, Proteins: Structure and Molecular Properties, W. H. Freeman & Co., San Francisco, 1983, pp 79-86), acetylation of the N-terminal amino group and amidation of any C-terminal carboxyl group.

Another type of covalent modification of antibody constructs that is within the scope of this invention involves altering the glycosylation profile of a protein. As is known in the art, glycosylation profiles can depend both on the sequence of the protein (eg, the presence or absence of specific glycosylation amino acid residues discussed below) and on the host cell or organism in which the protein is produced. Specific expression systems are discussed below.

Glycosylation of polypeptides is typically N-linked or O-linked. N-linked glycosylation refers to the attachment of a carbohydrate group to the side chain of an asparagine residue. The tripeptide sequences asparagine-X-serine and asparagine-X-threonine, where X is any amino acid except proline, represent recognition sequences for enzymatic attachment of a carbohydrate component to the asparagine side chain. Therefore, the presence of any of these tripeptide sequences creates a potential glycosylation site. O-linked glycosylation refers to the attachment of one of the sugars Nacetylgalactosamine, galactose or xylose to a hydroxyamino acid, most commonly serine or threonine, although 5-hydroxyproline or 5-hydroxylysine can also be used.

The addition of glycosylation sites to the disclosed antibody construct is conveniently accomplished by changing the amino acid sequence to include one or more of the tripeptide sequences described above (for N-linked glycosylation sites). The change can also be made by adding or substituting one or more serine or threonine residues in the original sequence (for O-linked glycosylation sites). For convenience, the amino acid sequence of the antibody construct is preferably changed at the DNA level, in particular by mutating the DNA encoding the target polypeptide at preselected bases so that the generated codons are translated into the required amino acids.

Another way to increase the number of carbohydrate components in an antibody construct is to chemically or enzymatically couple glycosides to a protein. These procedures are advantageous because they do not require the production of a protein in the host cell that is capable of N- or O-linked glycosylation. Depending on the coupling method used, the sugar (a) can be attached to (a) arginine and histidine, (b) free carboxyl groups, (c) free sulfhydryl groups such as cysteine sulfhydryl groups, (d) free hydroxyl groups such as as hydroxyl groups of serine, threonine or hydroxyproline, (e) aromatic residues such as aromatic residues of phenylalanine, tyrosine or tryptophan, or (f) glutamine amide group. These methods are described in WO 87/05330, and Aplin and Wriston, 1981, CRC Crit. Rev. Biochem., pp. 259-306.

Removal of carbohydrate residues present in the original antibody construct may be carried out chemically or enzymatically. For chemical deglycosylation, it is necessary to expose the protein to the action of a trifluoromethanesulfonic acid compound or an equivalent compound. This treatment results in the breakdown of most or all of the sugars except the linking sugar (N-acetylglucosamine or N-acetylgalactosamine) while leaving the polypeptide intact. Chemical deglycosylation is described in Hakimuddin et al., 1987, Arch. Biochem. Biophys. 259:52 and by Edge et al., 1981, Anal. Biochem. 118:131. Enzymatic cleavage of carbohydrate residues on polypeptides can be achieved using a range of endo- and exoglycosidases, as described in Thotakura et al., 1987, Meth. Enzymol. 138:350. Glycosylation at potential glycosylation sites can be prevented using the compound tunicamycin as described in Duskin et al., 1982, J. Biol. Chem. 257:3105. Tunikamycin blocks the formation of protein-N-glycoside bonds.

Other modifications to the antibody design are also discussed in this document. For example, another type of covalent modification of an antibody construct includes linking the antibody construct to various non-protein polymers, including, but not limited to, various polyols such as polyethylene glycol, polypropylene glycol, polyoxyalkylenes, or copolymers of polyethylene glycol and polypropylene glycol, in the manner described in US Pat. Nos. 4,640,835; 4496689; 4301144; 4670417; 4,791,192 or 4,179,337. Also, as is known in the art, amino acid substitutions can be made at various positions within an antibody construct to facilitate the addition of polymers such as PEG.

In some embodiments of the invention, covalent modification of antibody constructs of the invention includes the addition of one or more labels. The tag group can be coupled to the antibody construct via spacer legs of various lengths to reduce potential steric mismatch. Various methods of labeling proteins are known in the art and can be used to carry out the present invention. The term label or label

- 24 042775 This group means any detectable label. In general, tags are divided into different classes depending on the method of detection - the following examples include, but are not limited to:

a) isotope labels, which may be radioactive or heavy isotopes such as radioisotopes or radionuclides (e.g. ³ H, ¹⁴ C, ¹⁵ N, ³⁵ S, ⁸⁹ Zr, ⁹⁰ Y, ⁹⁹ Tc, ¹ⁿ In, ¹²⁵ I, ¹³¹ I );

b) magnetic marks (eg magnetic particles);

c) redox active groups;

d) optical dyes (including, but not limited to, chromophores, luminophores, and fluorophores) such as fluorescent groups (e.g., FITC, rhodamine, lanthanide phosphorus), chemiluminescent groups, and fluorophores, which can be either low molecular weight fluorophores or protein fluorophores ;

e) enzymatic groups (eg horseradish peroxidase, β-galactosidase, luciferase, alkaline phosphatase);

f) biotinylated groups;

g) predefined polypeptide epitopes recognized by the secondary reporter (eg, paired leucine zipper sequences, secondary antibody binding sites, metal binding domains, epitope tags, etc.).

The fluorescent label may be any molecule that can be detected due to its inherent fluorescent properties. Suitable fluorescent labels include, but are not limited to, fluorescein, rhodamine, tetramethylrhodamine, eosin, erythrosin, coumarin, methylcoumarins, pyrene, malachite green, stilbene, lucifer yellow, cascade blue J, Texas red, IAEDANS, EDANS, BODIPY FL, LC Red 640, Cy 5, Cy 5.5, LC Red 705, Oregon Green, Alexa-Fluor (Alexa Fluor 350, Alexa Fluor 430, Alexa Fluor 488, Alexa Fluor 546, Alexa Fluor 568, Alexa Fluor 594, Alexa Fluor 633, Alexa Fluor 660, Alexa Fluor 680), Cascade Blue, Cascade Yellow, and R-Phycoerythrin (PE) (Molecular Probs, Eugene, OR), FITC, Rhodamine, and Texas Red (Pierce, Rockford, IL), Cy5, Cy5.5, Cy7 (Amersham Life Science, Pittsburgh, PA). Suitable optical dyes, including fluorophores, are described in the Molecular Probes Handbook by Richard P. Haugland.

Suitable protein fluorescent labels also include, but are not limited to, green fluorescent protein, including GFP of Renilla, Ptilosarcus or Aequorea species (Chalfie et al., 1994, Science 263:802-805), EGFP (enhanced green fluorescent protein) (Clontech Laboratories , Inc., Genbank accession number U55762), blue fluorescent protein (BFP, Quantum Biotechnologies, Inc. 1801 de Maisonneuve Blvd. West, 8th Floor, Montreal, Quebec, Canada H3H 1J9; Stauber, 1998, Biotechniques 24:462-471; Heim et al., 1996, Curr. Biol. 6:178-182), enhanced yellow fluorescent protein (EYFP, Clontech Laboratories, Inc.), luciferase (Ichiki et al., 1993, J. Immunol. 150:5408-5417 ), β-galactosidase (Nolan et al., 1988, Proc. Natl. Acad. Sci. U.S.A. 85:2603-2607) and Renilla (WO 92/15673, WO 95/07463, WO 98/14605, WO 98/26277, WO 99/49019, US Pat. Nos. 5,292,658; 5,418,155; 5,683,888; 5,741,668;

Leucine zipper domains are peptides that promote oligomerization of the proteins in which they are found. Leucine zippers were originally identified in several DNA-binding proteins (Landschulz et al., 1988, Science 240:1759) and have since been found in a variety of different proteins. Among known leucine zippers are naturally occurring peptides and their derivatives that dimerize or trimerize. Examples of leucine zipper domains suitable for making soluble oligomeric proteins are described in PCT application WO 94/10308, and leucine zipper derived from lung surfactant protein D (SPD) is described in Hoppe et al., 1994, FEBS Letters 344:191. The use of a modified leucine zipper that provides stable trimerization of the heterologous protein fused to it is described in Fanslow et al., 1994, Semin. Immunol. 6:267-78. In one approach, recombinant fusion proteins containing the CDH3 antibody fragment or derivatives fused to the leucine zipper peptide are expressed in suitable host cells, and the soluble CDH3 oligomeric fragments or derivatives that form are isolated from the culture supernatant.

The antibody construct of the invention may also contain additional domains that are, for example, useful in isolating the molecule or relating to an adapted pharmacokinetic profile of the molecule. Domains useful for isolating an antibody construct can be selected from peptide motifs or re-introduced fragments that can be captured by an isolation technique such as an isolation column. Non-limiting embodiments of such additional domains include peptide motifs known as Myc-tag, HAT-tag, HA-tag, TAP-tag, GST-tag, chitin binding domain (CBD-tag), maltose-binding protein (MBC-tag), Flag -tag, Strep-tag and their variants (e.g. StrepII-tag) and His-tag. All antibody constructs described herein, characterized by identified CDRs, preferably containing a His-tag domain, which is commonly known as a repeat of consecutive His residues in the amino acid sequence of the molecule, preferably five, and more preferably six His residues (hexahistidine, see SEQ ID NO :

- 25 042775

436). The His-tag may be located, for example, at the N- or C-terminus of the antibody construct, preferably it is located at the C-terminus. Most preferably, the hexahistidine tag (HHHHHH) is linked via a peptide bond to the C-terminus of the antibody construct of the invention.

The first binding domain of the antibody construct of the invention binds to human CDH3 on the surface of the target cell. The amino acid sequence of human CDH3 is shown in SEQ ID NO: 1. It is understood that the term on the surface in the context of this invention means that the binding domain specifically binds to an epitope or epitope cassette contained in the extracellular domain of CDH3 (CDH3 ECD). The first binding domain of the invention therefore preferentially binds to CDH3 when it is expressed naturally by expressing cells or cell lines and/or cells or cell lines transformed or (stably/transiently) transfected with CDH3. In a preferred embodiment, the first binding domain also binds to CDH3 when CDH3 is used as a target molecule or ligand in an in vitro binding assay such as BIAcore or Scatchard. The target cell can be any prokaryotic or eukaryotic cell expressing CDH3 on its surface; preferably, the target cell is a cell that is part of the human or animal body, such as a tumor or cancer cell.

The term CDH3 ECD refers to a form of CDH3 that is substantially free of the transmembrane and cytoplasmic domains of CDH3. One of skill in the art will appreciate that the transmembrane domain identified for a CDH3 polypeptide of the invention is identified according to criteria commonly used in the art to identify this type of hydrophobic domain. The exact boundaries of the transmembrane domain may vary, but most likely by no more than 5 amino acids at either end of the domain specifically mentioned herein. The preferred human CDH3 ECD is shown in SEQ ID NO: 3.

The affinity of the first binding domain for human CDH3 is preferably <15 nM, more preferably <10 nM, even more preferably <5 nM, even more preferably <1 nM, even more preferably <0.5 nM, even more preferably <0.1 nM and most preferably <0.05 nM. Affinity can be measured, for example, in the analysis of BIAcore or in the analysis of Scatchard, for example, as described in the examples. Other methods for determining affinity are well known to those skilled in the art.

T cells or T lymphocytes are a type of lymphocyte (itself a type of white blood cell) that play a central role in cell-mediated immunity. There are several subsets of T cells, each with a specific function. T cells can be distinguished from other lymphocytes such as B cells and NK cells by the presence of a T cell receptor (TCR) on the cell surface. The TCR is responsible for the recognition of antigens associated with large histocompatible complex (MHC) molecules and is composed of two distinct protein chains. In 95% of T cells, the TCR consists of an alpha (a) and a beta (b) chain. When the TCR interacts with an antigenic peptide and an MHC (peptide/MHC complex), the T-lymphocyte is activated through a series of biochemical events mediated by associated enzymes, co-receptors, specialized adaptive molecules, and activated or released transcription factors.

The CD3 receptor complex is a protein complex and consists of four chains. In mammals, the complex contains a CD3y chain (gamma), a CD3δ chain (delta), and two CD3ε chains (epsilon). These chains bind to the T cell receptor (TCR) and the so-called ζ (zeta) chain to form the T cell CD3 receptor complex and generate an activation signal in T lymphocytes. The CD3y (gamma), CD3δ (delta) and CD3ε (epsilon) chains are tightly bound cell surface proteins of the immunoglobulin superfamily containing a single immunoglobulin extracellular domain. The intracellular tails of CD3 molecules contain one conserved motif, known as the immunoreceptor-based tyrosine-based activation motif, or short transient, which is required for TCR signaling. The CD3 epsilon molecule is a polypeptide that in humans is encoded by the CD3E gene, which is located on chromosome 11. The most preferred CD3 binding epitope corresponds to amino acid residues 1-27 of the extracellular domain of human CD3 epsilon.

Redirected lysis of target cells by T cell recruitment with a multispecific, at least bispecific, antibody construct involves the formation of a cytolytic synapse and the delivery of perforin and granzymes. Engaged T cells are capable of serial lysis of target cells and are not affected by immune evacuation mechanisms that interfere with peptide antigen processing and presentation or clonal T cell differentiation; see, for example, WO 2007/042261.

Cytotoxicity mediated by CDH3/CD3 bispecific antibody constructs can be measured in various ways. The effector cells can be, for example, stimulated enriched (human) CD8 positive T cells or unstimulated (human) peripheral blood mononuclear cells (PBMC). If the target cells are

- 26 042775 macaque or macaque CDH3 is expressed or transfected, the effector cells must also be of macaque origin, such as a macaque T cell line, eg 4119LnPx. Target cells must express (at least the extracellular domain of) CDH3, eg human or macaque CDH3. The target cells can be a cell line (such as CHO) that is stably or transiently transfected with CDH3, eg human or macaque CDH3. Alternatively, the target cells may be a cell line naturally expressing CDH3. Generally, EC ₅₀ values are expected to be lower with target cell lines expressing higher levels of CDH3 on the cell surface. The target cell effector ratio (E:T) is typically around 10:1, but may also vary. Cytotoxic activity of CDH3/CD3 bispecific antibody constructs can be measured in a 51-chromium release assay (incubation time about 18 hours) or in a FACS-based cytotoxicity assay (incubation time about 48 hours). Modifications to the assay incubation time (cytotoxic reaction) are also possible. Other methods for measuring cytotoxicity are well known to one of skill in the art and include MTT or MTC assays, ATP based assays including bioluminescent assays, sulpopodamine B (SRB) assay, WST assay, clonogenic assay, and ECIS technology.

Cytotoxic activity mediated by the CDH3/CD3 bispecific antibody constructs of the invention is preferably measured in a cell-based cytotoxicity assay. It can also be measured in a 51-chromium release assay. It is represented by the EC ₅₀ value, which corresponds to half the maximum effective concentration (the concentration of the antibody construct that induces a cytotoxic response halfway between baseline and maximum). Preferably, the EC ₅₀ value of the CDH3xCD3 bispecific antibody constructs is <5000 pg/ml or <4000 pg/ml, more preferably <3000 pg/ml or <2000 pg/ml, even more preferably < 1000 pg/ml or <500 pg/ml , even more preferably <400 pg/ml or <300 pg/ml, even more preferably <200 pg/ml, even more preferably <100 pg/ml, even more preferably <50 pg/ml, even more preferably <20 pg/ml ml or <10 pg/ml and most preferably <5 pg/ml.

The above EC ₅₀ values can be measured in different tests. One skilled in the art will recognize that the EC ₅₀ value is expected to be lower when stimulated/enriched CD8+ T cells are used as effector cells compared to unstimulated PBMC. Furthermore, EC50 values can be expected to be lower when the target cells express a large amount of the target antigen compared to the rat with restricted target expression. For example, when stimulated/enriched human CD8+ T cells are used as effector cells (and CDH3-transfected cells such as CHO cells or a CDH3-positive naturally expressing cell line such as A431 are also used as target cells), the EC ₅₀ value The CDH3xCD3 of the bispecific antibody construct is preferably < 1000 pg/ml, more preferably < 500 pg/ml, even more preferably < 250 pg/ml, even more preferably < 100 pg/ml, even more preferably < 50 pg/ml, even more preferably <10 pg/ml and most preferably <5 pg/ml. When human PBMCs are used as effector cells, the _EC50 value of the CDH3xCD3 bispecific antibody construct is preferably <5000 pg/mL or <4000 pg/mL (particularly when the target cells are a naturally expressing CDH3 positive cell line such as A431) , more preferably <2000 pg/ml (particularly when the target cells are CDH3 transfected cells such as CHO cells), more preferably <1000 pg/ml or <500 pg/ml, even more preferably <200 pg/ml, even more preferably <150 μg/ml, even more preferably <100 pg/ml and most preferably <50 pg/ml or less. When a macaque T cell line such as LnPx4119 is used as effector cells and a transfected macaque CDH3 cell line such as CHO cells is used as target cell line, the EC50 value of the CDH3xCD3 bispecific antibody construct is preferably <2000 pg/ml or <1500 pg/ml, more preferably <1000 pg/ml or <500 pg/ml, even more preferably <300 pg/ml or <250 pg/ml, even more preferably <100 pg/ml and most preferably <50 pg /ml

Preferably, the CDH3/CD3 bispecific antibody constructs of the invention do not induce/mediate lysis or substantially induce/mediate lysis of CDH3-negative cells such as CHO cells. The term does not induce lysis, does not substantially cause lysis, does not mediate lysis, or does not substantially mediate lysis, means that the antibody construct of the invention does not induce or mediate lysis by no more than 30%, preferably no more than 20%, more preferably not more than 10%, particularly preferably not more than 9%, 8%, 7%, 6% or 5% CDH3 negative cells, resulting in 100% lysis of a CDH3 cell line such as A431. This is typically applied for antibody construct concentrations up to 500 nM. One skilled in the art knows how to measure cell lysis without additional steps. In addition, this document provides specific instructions for measuring cell lysis.

The difference in cytotoxic activity between monomeric and dimeric isoforms of individual constructs of CDH3/CD3-bispecific antibodies is called specific activity gap.

- 27 042775 sti. This gap in specific activity can, for example, be calculated as the ratio between the _EC50 values of the monomeric and dimeric form of the molecule. The specific activity gaps of the CDH3/CD3 bispecific antibody constructs of this invention are preferably < 5, more preferably < 4, even more preferably < 3, even more preferably < 2, and most preferably < 1. As an example, the specific activity gap for CDH3-13 was determined to be 0.5, the specific activity gap for CDH3-13xCD3-HLE (Fc) was determined to be 0.9, the specific activity gap for CDH3-25 was 0.7, and the specific activity gap for CDH3-25xCD3-HALB was also defined as 0.7.

The first and/or second (or any other) binding domain(s) of an antibody construct of the invention is/are preferably cross-species specific for the mammalian primate order. Cross-species specific CD3 binding domains are, for example, described in WO 2008/119567. In one embodiment, the first and/or second binding domain, in addition to binding to human CDH3 and human CD3, respectively, will also bind to primate CDH3/CD3, including (but not limited to) New World primates (such as Callithrix jacchus, Saguimis Oedipus or Saimiri sciureus), Old World primates (such as baboons and macaques), gibbons, and non-human hominids.

In one aspect of the invention, the first binding domain binds to human CDH3 and further binds to macaque CDH3, such as Macaca fascicularis CDH3 (SEQ ID NO: 5), and more preferably macaque ECD CDH3. The affinity of the first binding domain for macaque CDH3 is preferably <15 nM, more preferably <10 nM, even more preferably <5 nM, even more preferably <1 nM, even more preferably <0.5 nM, even more preferably <0.1 nM and most preferably <0.05 nM or even <0.01 nM.

Preferably, the affinity gap of the antibody constructs of the invention for binding macaque CDH3 versus human CDH3 [ma CDH3:hu CDH3] (as determined, for example, by the BiaCore or Scatchard method, see examples) is 0.1 to 10, more preferably 0 .2 to 5, even more preferably 0.3 to 2.5, even more preferably 0.4 to 2, and most preferably 0.5 to 1.

In one embodiment of the antibody construct of the invention, the second binding domain binds to human CD3 epsilon and Callithrix jacchus, Saguimis Oedipus, or Saimiri sciureus. Preferably, the second binding domain binds to the extracellular epitope of these CD3 epsilon chains. The second binding domain is also expected to bind to an extracellular epitope of the CD3 epsilon chain of human and Macac. The most preferred CD3 epsilon epitope consists of amino acid residues 1-27 of the extracellular domain of human CD3 epsilon. More specifically, the epitope contains at least the amino acid sequence Gln-Asp-Gly-Asn-Glu. Callithrix jacchus and Saguinus oedipus are New World primates belonging to the family Callitrichidae while Saimiri sciureus is a New World primate belonging to the family Cebidae.

It is particularly preferred for the antibody construct of this invention that the second binding domain that binds to human CD3 on the surface of the T cell contains a VL region containing CDR-L1, CDR-L2 and CDR-L3 selected from:

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

(b) CDR-L1 as shown in SEQ ID NO: 117 WO 2008/119567, CDR-L2 as shown in SEQ ID NO: 118 WO 2008/119567 and CDR-L3 as shown in SEQ ID NO: 119 WO 2008/119567; and (c) CDR-L1 as shown in SEQ ID NO: 153 of WO 2008/119567, CDR-L2 as shown in SEQ ID NO: 154 of WO 2008/119567 and CDR-L3 as shown in SEQ ID NO: 155 WO2008/119567;

In an alternative preferred embodiment of the antibody construct of this invention, the second binding domain that binds to human CD3 on the surface of the T cell comprises a VH region containing CDR-H1, CDR-H2 and CDR-H3 selected from:

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

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

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

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

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

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

(g) CDR-H1 as shown in SEQ ID NO: 120 WO 2008/119567, CDR-H2 as shown in SEQ ID NO:

- 28 042775

121 WO 2008/119567 and CDR-H3 as shown in SEQ ID NO: 122 WO 2008/119567;

(h) CDR-H1 as shown in SEQ ID NO: 138 of WO 2008/119567, CDR-H2 as shown in SEQ ID NO: 139 of WO 2008/119567, and CDR-H3 as shown in SEQ ID NO: 140 WO2008/119567;

(i) CDR-H1 as shown in SEQ ID NO: 156 WO 2008/119567, CDR-H2 as shown in SEQ ID NO: 157 WO 2008/119567 and CDR-H3 as shown in SEQ ID NO: 158 WO 2008/119567; and (j) CDR-H1 as shown in SEQ ID NO: 174 of WO 2008/119567, CDR-H2 as shown in SEQ ID NO: 175 of WO 2008/119567 and CDR-H3 as shown in SEQ ID NO: 176 WO 2008/119567.

It is also preferred for the antibody construct of this invention that the second binding domain that binds to human CD3 on the surface of the T cell contains a VL region selected from the group consisting of a VL region as shown in SEQ ID NOs: 35, 39, 125 , 129, 161 or 165 of WO 2008/119567.

Alternatively, it is preferred that the second binding domain that binds to human CD3 on the surface of the T cell contains a VH region selected from the group consisting of a VH region as shown in SEQ ID NOs: 15, 19, 33, 37, 51, 55 , 69, 73, 87, 91, 105, 109, 123, 127, 141, 145, 159, 163, 177 or 181 of WO 2008/119567.

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

(a) a VL region as shown in SEQ ID NO: 17 or 21 of WO 2008/119567 and a VH region as shown in SEQ ID NO: 15 or 19 of WO 2008/119567;

(b) a VL region as shown in SEQ ID NO: 35 or 39 of WO 2008/119567 and a VH region as shown in SEQ ID NO: 33 or 37 of WO 2008/119567;

(c) a VL region as shown in SEQ ID NO: 53 or 57 of WO 2008/119567 and a VH region as shown in SEQ ID NO: 51 or 55 of WO 2008/119567;

(d) a VL region as shown in SEQ ID NO: 71 or 75 of WO 2008/119567 and a VH region as shown in SEQ ID NO: 69 or 73 of WO 2008/119567;

(e) a VL region as shown in SEQ ID NO: 89 or 93 of WO 2008/119567 and a VH region as shown in SEQ ID NO: 87 or 91 of WO 2008/119567;

(f) a VL region as shown in SEQ ID NO: 107 or 111 of WO 2008/119567 and a VH region as shown in SEQ ID NO: 105 or 109 of WO 2008/119567;

(g) a VL region as shown in SEQ ID NO: 125 or 129 of WO 2008/119567 and a VH region as shown in SEQ ID NO: 123 or 127 of WO 2008/119567;

(h) a VL region as shown in SEQ ID NO: 143 or 147 of WO 2008/119567 and a VH region as shown in SEQ ID NO: 141 or 145 of WO 2008/119567;

(i) a VL region as shown in SEQ ID NO: 161 or 165 of WO 2008/119567 and a VH region as shown in SEQ ID NO: 159 or 163 of WO 2008/119567; and (j) a VL region as shown in SEQ ID NO: 179 or 183 of WO 2008/119567 and a VH region as shown in SEQ ID NO: 177 or 181 of WO 2008/119567.

The above binding domains that bind to human CD3 and are described in WO 2008/119567 are also shown in the present SEQ ID NOs: 445-537.

In a preferred embodiment of the antibody design of the invention, the binding domains, and in particular the second binding domain (which binds to human CD3 on the surface of a T cell), are in the following format: The pairs of VH regions and VL regions are in single chain antibody (scFv) format. The VH and VL sections are arranged in the order VH-VL or VLVH. Preferably, the VH region is positioned N-terminally to the linker sequence and the VL region is positioned from the C-terminus of the linker sequence.

A preferred embodiment of the above antibody construct of the present invention is characterized by a second binding domain that binds to human CD3 on the surface of a T cell containing 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 of WO 2008/119567.

In one embodiment of the present invention, the antibody construct has an amino acid sequence selected from the group consisting of those sequences shown in SEQ ID NO: 158, SEQ ID NO: 168, SEQ ID NO: 178, SEQ ID NO: 188, SEQ ID NO: 198, SEQ ID NO: 208, SEQ ID NO: 218, SEQ ID NO: 228, SEQ ID NO: 238 and SEQ ID NO: 248.

In another embodiment of the present invention, the antibody construct has an amino acid sequence selected from the group consisting of those sequences shown in SEQ ID NO: 379, SEQ ID NO: 380, SEQ ID NO: 381, SEQ ID NO: 382, SEQ ID NO: 383, SEQ ID NO: 384, SEQ ID NO: 385, SEQ ID NO: 386, SEQ ID NO: 387, SEQ ID NO: 388 and SEQ ID NO: 389.

In another embodiment of the present invention, the antibody construct has an amino acid sequence selected from the group consisting of those sequences shown in SEQ ID NO: 288, SEQ ID NO: 298, SEQ ID NO: 308, SEQ ID NO: 318, SEQ ID NO: 328, SEQ ID NO:

- 29 042775

338, SEQ ID NO: 348 and SEQ ID NO: 358.

Modifications to the amino acid sequences of the antibody constructs described herein are also contemplated. For example, it may be desirable to improve the affinity and/or other biological properties of an antibody construct. Amino acid sequence variants of antibody constructs are obtained by making appropriate nucleotide modifications to the nucleotide sequence encoding the antibody construct, or by peptide synthesis. All amino acid sequence modifications described below should result in an antibody construct that still retains the desired biological activity (binding to CDH3 and CD3) of the unmodified parent molecule.

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

Amino acid modifications include, for example, deletions and/or insertions and/or substitutions of residues in the amino acid sequences of an antibody construct. Any combination of deletions, insertions, and substitutions can be made to produce the final construct, as long as the final construct has the desired characteristics. Amino acid substitutions can alter the post-translational processes of antibody construction, such as changing the number or position of glycosylation sites.

For example, 1, 2, 3, 4, 5, or 6 amino acids can be inserted or deleted in each of the CDRs (depending on their length, of course), while 1, 2, 3, 4, 5, 6, 7, 8 , 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or 25 amino acids may be inserted or deleted in each of the FRs. Preferably inserts in the amino acid sequence include N- and/or C-terminal hybrids from 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10 residues in length to polypeptides containing one hundred or more residues, as well as inserts of one or multiple amino acid residues within a sequence. An insert variant of an antibody construct of the invention comprises fusion to the N-terminus or C-terminus of the enzyme antibody construct, or fusion to a polypeptide that increases the serum half-life of the antibody construct.

Sites of greatest interest for displacement mutagenesis include heavy and/or light chain CDRs, in particular hypervariable regions, but light and/or heavy chain FR rearrangements are also contemplated. The substitutions are preferably conservative substitutions as described herein. Preferably 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10 amino acids may be substituted in the CDR, while 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14, 15, 16, 17, 18, 19, 20, or 25 amino acids may be substituted in the framework regions (FR), depending on the length of the CDR or FR. For example, if a CDR sequence contains 6 amino acids, one, two, or three of those amino acids are assumed to be substituted. Similarly, if the CDR sequence contains 15 amino acids, one, two, three, four, five, or six of these amino acids are assumed to be substituted.

A useful method for identifying certain residues or regions of antibody constructs that are preferred sites for mutagenesis is called alanine scanning mutagenesis, as described by Cunningham and Wells in Science, 244: 1081-1085 (1989). This method identifies a residue or group of target residues within an antibody construct (e.g., charged residues such as arg, asp, his, lys, and glu) and replaces them with neutral or negatively charged amino acids (most preferably alanine or polyalanine) to achieve an interaction amino acids with an epitope.

Such amino acid positions that are functionally sensitive to substitutions are further modified by introducing additional or other variants at the substitution sites. Thus, while the site or site of introduction of amino acid sequence variants is predetermined, the nature of the mutation per se may not necessarily be predetermined. For example, to analyze or optimize the occurrence of a mutation at a given site, alanine-scanning or selective mutagenesis can be performed at the target codon and target region and expressed antibody variants are screened for the optimal combination of a given activity. Methods for introducing substitution mutations at predetermined sites in DNA having a known sequence are well known, such as M13 primer mutagenesis and PCR mutagenesis. Mutant screening is performed using antigen binding activity assays, such as the CDH3 binding assay or

- 30 042775

CD3.

In general, if amino acids are substituted in one or more or all of the heavy and/or light chain CDRs, it is preferred that the substituted sequence then obtained is at least 60%, more preferably 65%, even more preferably 70%, particularly preferably 75 %, more preferably 80% identical to the original CDR sequence. This means that it depends on the length of the CDR to what extent it is identical to the substituted sequence. For example, a CDR containing 5 amino acids is preferably 80% identical to its substituted sequence in order to have at least one substituted amino acid. Accordingly, the CDRs of an antibody construct may have varying degrees of identity with their substituted sequences, for example, CDRL1 may have 80% while CDRL3 may have 90%.

Preferred substitutions (or substitutions) are conservative substitutions. However, any substitution (including non-conservative substitution or one or more of the exemplary substitutions shown in Table 1 below) is contemplated as long as the antibody construct retains its ability to bind through the first binding domain and to CD3 or CD3 epsilon through the second binding domain and/ or its CDRs are at least 60%, more preferably 65%, even more preferably 70%, more preferably 75%, most preferably 80% identical to the original CDR sequence with the subsequently substituted sequence.

Conservative substitutions are given in table. 1 under the heading preferred substitutions. If such substitutions change the biological activity, then they are the most significant changes and are given in table. 1 under the heading generic substitutions, or as further described below with respect to amino acid classes, can be introduced and products screened.

Table 1. Amino acid substitutions

Initial Typical substitutions Preferred substitutions Ala (A) val, leu, ile val Arg(R) lys, gin, asn lys Asn(N) gin, his, asp, lys, arg gin Asp(D) glu, asn glu Cys(C) ser, ala ser Gin (Q) asn, glu asn Glu(E) asp, gin asp Gly(G) Ala ala His(H) asn, gin, lys, arg arg He(I) leu, val, met, ala, phe leu Leu(L) norleucine, ile, val, met, ala ile Lys (K) arg, gin, asn arg Met(M) leu, phe, ile leu Phe(F) leu, val, ile, ala, tyr tyr Pro (P) Ala ala Ser(S) Thr thr Thr(T) Ser ser TRP(W) tyr, phe tyr Tyr(Y) trp, phe, thr, ser phe Val(V) ile, leu, met, phe, ala leu

Significant modifications to the biological properties of the antibody constructs of this invention are accompanied by selective substitutions that differ significantly in their ability to maintain: (a) the structure of the polypeptide backbone in the region of substitution, such as a folded or helical conformation, (b) the charge or hydrophobicity of the molecule at the target site, and (c ) side chain dimensions. Residues of natural origin are divided into groups depending on the general properties of the side chains: (1) hydrophobic: norleucine, met, ala, val, leu, ile; (2) neutral hydrophilic: cys, ser, thr; (3) acidic: asp, glu; (4) basic: asn, gin, his, lys, arg; (5) residues that affect chain orientation: gly, pro; and (6) aromatic: trp, tyr, phe.

Non-conservative substitutions lead to the replacement of a representative of one of these classes by a representative of another class. Cysteine residues that are not involved in maintaining the proper conformation of the antibody construct can also be replaced primarily with serine to enhance the oxidative stability of the molecule and prevent unwanted crosslinking. Conversely, cysteine bonds can be introduced into an antibody molecule to improve its stability (especially when the antibody is an antibody fragment, such as an Fv fragment).

In the case of amino acid sequences, sequence identity and/or similarity is determined using standard methods known in the art, including, but not limited to, the Local Sequence Identity Algorithm by Smith and Waterman, 1981, Adv. Appl. Math. 2:482, sequence identity alignment algorithm Needleman and Wunsch, 1970, J. Mol. Biol. 48:443, similarity search method Pearson and Lipman, 1988, Proc. Nat. Acad. sci. U.S.A. 85:2444, computerized implementations of these algorithms (GAP, BESTFIT, FASTA, and TFASTA in the Wisconsin Genetics software package, Genetics Computer Group, 575 Science Dr., Madison, WI), Best Fit sequence program described by Devereux et al., 1984, Nucl. Acid Res. 12:387-395, preferably used with the default parameters or fitting method. Preferably, percent identity is calculated by FastDB based on the following parameters: mismatch penalty 1; penalty for gap 1; gap size penalty 0.33; and compound penalty 30, Current Methods in Sequence Comparison and Analysis, Macromolecule Sequencing and Synthesis, Selected Methods and Applications, pp. 127-149 (1988), Alan R. Liss, Inc.

An example of an algorithm used is PILEUP. PILEUP creates a multiple sequence alignment from a group of related sequences using a progressive, pairwise alignment. It is also possible to build a tree showing the group relationships used to create the alignment. PILEUP uses a simplification of the progressive leveling method Feng & Doolittle, 1987, J. Mol. Evol. 35:351-360; the method is similar to that described in Higgins and Sharp, 1989, CABIOS 5:151-153. Applied PILEUP parameters include a default gap weight of 300, a default gap length weight of 0.10, and weighted end gaps.

Another example of a useful algorithm is the BLAST algorithm described in Altschul et al., 1990, J. Mol. Biol. 215:403-410; Altschul et al., 1997, Nucleic Acids Res. 25:3389-3402 and Karin et al., 1993, Proc. Natl. Acad. sci. U.S.A. 90:5873-5787. A particularly suitable BLAST program is the WU-BLAST-2 program, which was obtained from Altschul et al., 1996, Methods in Enzymology 266:460-480. WUBLAST-2 uses several search options, most of which have default values set. Adjustable parameters are set with the following values for proteins: overlap length = 1, overlap fraction = 0.125, threshold segment length, T = 11. Parameters HSP S and HSP S2 are dynamic values and are set by the program itself depending on the composition of a particular sequence and the composition of a particular base the data against which the sequence of interest is searched; however, these values can be corrected to increase the sensitivity.

An additional applicable algorithm is gapped BLAST, described in Altschul et al., 1993, Nucl. Acids Res. 25:3389-3402. Gap BLAST uses the BLOSUM-62 substitution weighting matrix; the threshold parameter T is set to 9; two-match method for triggering extensions without gaps, cost per gap length to cost 10+k; Xu is set to 16 and Xg is set to 40 in the database search step and 67 in the output step of the algorithms. Gap alignment is triggered at an estimate corresponding to approximately 22 bits.

In general, the amino acid homology, similarity, or identity between individual variant CDRs is at least 60% relative to the sequences illustrated herein, and generally with preferably increasing homology or identity of at least 65% or 70%, more preferably at least 75% or 80%, most preferably at least 85%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, and substantially 100%. Likewise, percent (%) nucleotide sequence identity relative to the nucleotide sequence of the binding proteins identified herein is presented as the percentage of nucleotide residues in a candidate sequence that are identical to nucleotide residues in an antibody coding sequence. The specific method uses a WU-BLAST-2 BLASTN module with default parameters, with overlap length and overlap ratio set to 1 and 0.125, respectively.

In general, the nucleotide homology, similarity, or identity between the nucleotide sequences encoding individual variant CDRs and the nucleotide sequences depicted herein is at least 60% or more typically, with preferred increasing homologies or identities being at least 65, 70 , 75, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99 and almost 100%. Thus, a CDR variant is one with specified homology, similarity, or identity to the parent CDR of the invention and shares a biological function, including but not limited to at least 60, 65, 70, 75, 80, 81, 82, 83 , 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% specificity and/or activity of the parental CDR.

- 32 042775

In one embodiment, the bispecific antibody constructs of the invention exhibit high monomer yields under standard study scale conditions, eg, in a standard two-step purification process. Preferably, the monomer yield of the antibody constructs of the invention is > 0.25 mg/L supernatant, more preferably > 0.5 mg/L, even more preferably > 1 mg/L, and most preferably > 3 mg/L supernatant.

Similarly, the yield of isomers of dimeric antibody constructs, and hence the percentage of monomers of antibody constructs, can be determined (ie, monomer: (monomer+dimer)). The performance of the monomeric and dimeric antibody constructs and the calculated percentage of monomer can, for example, be obtained from the purification step of the SEC obtained from the culture supernatant from a standardized spinning flask manufacturing process. In one embodiment, the percentage of antibody construct monomers is >80%, more preferably >85%, even more preferably >90%, and most preferably >95%.

In yet another embodiment, the percent identity of the human germline antibody constructs of the invention is >70% or >75%, more preferably >80% or >85%, even more preferably >90%, and most preferably >95% (See Example 7). Identity with human germline antibody gene products is believed to be an important feature to reduce the risk of using therapeutic proteins to induce an immune response against a drug in a patient during treatment. Hwang and Foote (Immunogenicity of engineered antibodies; Methods 36 (2005) 3-10) demonstrated that reducing the number of non-human portions of antibody drug constructs resulted in a reduced risk of inducing anti-drug antibodies in patients during treatment. Comparing an exhaustive number of clinically assessed antibodies with corresponding immunogenicity data, it is shown that humanization of antibody V regions renders the protein less immunogenic (mean 5.1% of patients) than antibodies bearing intact non-human V regions (mean 23.59% patients). Therefore, a higher degree of identity with human sequences is desirable for V-region based protein therapy in the form of antibody constructs. For this purpose of determining germline identity, VL V regions can be aligned with the amino acid sequences of human germline V and J segments (http://vbase.mrc-cpe.cam.ac.uk/) using Vector NTI software. and amino acid sequence calculated by dividing identical amino acid residues by the total number of VL amino acid residues in percent. The same can be done for VH segments (http://vbase.mrc-cpe.cam.ac.uk/) except that VH CDR3 may be excluded due to its great diversity and lack of existing VH CDR3 alignment patterns human germ line. Recombinant techniques can then be used to increase the sequence identity of human germline genes.

In one embodiment, the antibody design has a preferred plasma stability (ratio of EC ₅₀ with plasma to EC ₅₀ without plasma) < 5, more preferably < 4 or < 3.5, even more preferably < 3 or < 2.5, and most preferably < 2 or < 1.5 or < 1. Plasma stability of the antibody construct can be tested by incubating the construct in human plasma at 37° C. for 24 h followed by determination of the EC ₅₀ in a cytotoxicity assay with 51-chromium release. Effector cells in the cytotoxicity assay can stimulate enriched human CD8-positive T cells. Target cells can, for example, be CHO cells transfected with human CDH3. The target cell effector ratio (E:T) can be selected as 10:1. The pool of human plasma used for this purpose is derived from the blood of healthy donors collected with EDTA-coated syringes. Cellular components are removed by centrifugation, and the upper phase of the plasma is collected and then combined. As a construct control, antibodies are diluted immediately prior to cytotoxicity assay in RPMI-1640 medium. Plasma stability is calculated as the ratio of EC ₅₀ (after plasma incubation) to EC ₅₀ (control) (See Example 11).

Preferably, the monomer-dimer conversion of the antibody constructs of the invention is low. The conversion can be measured under various conditions and analyzed using high performance size exclusion chromatography. For example, incubation of monomeric isoforms of antibody constructs can be carried out for 7 days at 37° C. and a concentration of, for example, 100 μg/ml or 250 μg/ml in an incubator. Under these conditions, it is preferred that the antibody constructs of the invention exhibit a dimer percentage of <5%, more preferably <4%, even more preferably <3%, even more preferably <2.5%, even more preferably <2%, still more preferably <1.5% and most preferably <1% (See example 9).

It is also preferred that the bispecific antibody constructs of the invention have very low dimeric conversion after a number of freeze/thaw cycles. For example, a monomeric antibody construct is brought to a concentration of 250 μg/mL in original formulation buffer and subjected to three freeze/thaw cycles (freeze at 80°C for 30 min followed by 30 min thaw at room temperature), and

- 33 042775 then high throughput SEC to determine the percentage of the original monomeric antibody construct that has been converted to a dimeric antibody construct. Preferably, the dimer percentages of the bispecific antibody constructs are <5%, more preferably <4%, even more preferably <3%, even more preferably <2.5%, even more preferably <2%, even more preferably <1.5%, and most preferably <1%, eg after three freeze/thaw cycles.

The bispecific antibody constructs of this invention preferably exhibit favorable thermal stability with aggregation temperatures above 50°C or above 52°C, more preferably above 54°C or above 55°C, even more preferably above 56°C or above 57°C, and most preferably above 58°C or above 59°C. The thermal stability parameter can be determined in terms of the antibody aggregation temperature as follows: an antibody solution at a concentration of 250 μg/ml is transferred into a disposable cuvette and placed in a dynamic light scattering device. The sample is heated from 40 to 70° C. at a heating rate of 0.5° C./min, continuously obtaining a measured radius. To calculate the aggregation temperature of an antibody, an increase in radius indicating protein melting and aggregation is used (See Example 10).

Alternatively, temperature melt curves can be determined using differential scanning calorimetry (DSC) to determine the intrinsic biophysical stability of antibody construct proteins. These experiments are performed using a VP-DSC MicroCal LLC device (Northampton, Massachusetts, USA). The energy capture of the sample containing the antibody construct is recorded from 20 to 90°C compared to the sample containing only drug buffer. The antibody constructs are adjusted to a final concentration of 250 μg/ml, eg in SEC buffer. To record an appropriate melting curve, the total temperature of the sample is increased in steps. At each temperature, the energy absorption of the sample and drug buffer is recorded. The difference in energy absorption Cp (kcal/mol/°C) of the sample minus the reference value is plotted against the respective temperature. The melting point is defined as the temperature at the first energy absorption maximum.

In addition, it is expected that CDH3xCD3 bispecific antibodies of the invention do not cross-react with (eg, do not bind to) the human CDH3 paralogs CDH1, CDH2, CDH4 and CDH5. In addition, it is expected that the CDH3xCD3 bispecific antibodies of the invention do not cross-react with (eg, do not bind to) the macaque/cynomolgus monkey CDH3 paralogs CDH1, CDH2, CDH4 and CDH5. See Example 6. The CDH3xCD3 bispecific antibodies of the invention also provide for a turbidity (measured by OD340 after adjusting the concentration of purified monomeric antibody to 2.5 mg/mL and overnight incubation) < 0.1, most preferably < 0.05 (See Example 12 ).

In yet another embodiment, the antibody construct of the invention is stable at acidic pH. The more tolerant the antibody construct behaves at a non-physiological pH, such as pH 5.5 (the pH required to run, for example, cation exchange chromatography), the greater the recovery of the antibody construct eluted from the ion exchange column relative to the total amount of protein loaded. Recovery of antibody construct from an ionic (e.g., cationic) exchange column at pH 5.5 is preferably >30%, more preferably >40%, more preferably >50%, even more preferably >60%, even more preferably >70%, still more preferably > 80% and most preferably > 90%.

Furthermore, the bispecific antibody constructs of this invention are expected to exhibit therapeutic efficacy or antitumor activity. This can, for example, be assessed in a study as described in Example 14. One of skill in the art knows how to modify or tailor certain parameters of this study, such as the number of tumor cells injected, the site of injection, the number of transplanted human T cells, the number of bispecific designs of antibodies to be administered, and timing, while achieving a meaningful and reproducible result. Preferably, the tumor growth inhibition T/C [%] is < 70 or < 60, more preferably < 50 or < 40, even more preferably < 30 or < 20, and most preferably < 10 or < 5 or even < 2.5.

The invention also provides a polynucleotide/nucleic acid molecule encoding an antibody construct of the invention.

A polynucleotide is a biopolymer consisting of 13 or more nucleotide monomers covalently linked in a chain. DNA (such as cDNA) and RNA (such as mRNA) are examples of polynucleotides with a distinct biological function. Nucleotides are organic molecules that serve as monomers or subunits of nucleic acid molecules such as DNA or RNA. The nucleic acid molecule or polynucleotide can be double-stranded and single-stranded, linear and cyclic. It is preferably contained in a vector, which is preferably contained in the host cell. The specified host cell, for example, after transformation or transfection with a vector or a polynucleotide according to the invention, is capable of expressing the const

- 34 042775 antibody production. For this purpose, the polynucleotide or nucleic acid molecule is operably linked to control sequences.

The genetic code is a set of rules by which information encoded in genetic material (nucleic acids) is translated into proteins. Biological decoding in living cells is carried out by the ribosome, which links the amino acids in the order determined by the mRNA, using tRNA molecules to carry the amino acids and read three nucleotides of the mRNA at a time. The code specifies how the sequences of these nucleotide triplets, called codons, determine which amino acid will be added next during protein synthesis. With few exceptions, a three-nucleotide codon in a nucleic acid sequence specifies one amino acid. Since the vast majority of genes are encoded with exactly the same code, this particular code is often referred to as the canonical or standard genetic code. Although the genetic code determines the protein sequence for a given coding region, other genomic regions can influence when and where these proteins are produced.

In addition, the present invention provides a vector containing the polynucleotide/nucleic acid molecule of the invention.

A vector is a nucleic acid molecule used as a carrier to transfer (foreign) genetic material into a cell. The term vector includes, but is not limited to, plasmids, viruses, cosmids, and artificial chromosomes. In general, the constructed vectors contain an origin of replication, a multiple cloning site, and a selectable marker. The vector itself is typically a nucleotide sequence, usually a DNA sequence, that contains an insert (transgene) and a larger sequence that serves as the backbone of the vector. Modern vectors may contain additional features besides the transgene insert and scaffold: a promoter, a genetic marker, antibiotic resistance, a reporter gene, a targeting sequence, a protein purification tag. Vectors, referred to as expression vectors (expression constructs), are specifically designed to express a transgene in a target cell and typically have control sequences.

The term control sequences refers to DNA sequences necessary for the expression of an operably linked coding sequence in a particular host organism. Control sequences that are suitable for prokaryotes include, for example, a promoter, an optional operator sequence, and a ribosome binding site. Promoters, polyadenylation signals, and enhancers are known to be used in eukaryotic cells.

A nucleic acid is operably linked when it is placed in a functional dependency of another nucleic acid sequence. For example, a presequence or secretory leader DNA is operably linked to a polypeptide's DNA if it is expressed as a precursor protein that participates in the secretion of the polypeptide; a promoter or enhancer is operably linked to a coding sequence if it affects the transcription of the sequence; or a ribosome binding site is operably linked to a coding sequence if it is located in such a way as to facilitate translation. In general, operably linked means that the DNA sequences, when linked, are contiguous and, in the case of a secretory leader, contiguous and in the reading phase. However, enhancers do not need to be contiguous. The bond is formed by ligation at suitable restriction sites. If such sites are not present, synthetic oligonucleotide adapters or linkers are used, in accordance with common practice.

Transfection is the process of intentionally introducing nucleic acid molecules or polynucleotides (including vectors) into target cells. This term is mainly used for non-viral methods in eukaryotic cells. Transduction is often used to describe the virus-mediated transfer of nucleic acid molecules or polynucleotides. Transfection of animal cells typically involves opening transient pores or holes in the cell membrane to allow uptake of the material. Transfection can be carried out using calcium phosphate, by electroporation, by squeezing cells, or by mixing a cationic lipid with material to produce liposomes that fuse with the cell membrane and deposit their contents inside.

The term transformation is used to describe the non-viral transfer of nucleic acid molecules or polynucleotides (including vectors) into bacteria as well as non-animal eukaryotic cells, including plant cells. Thus, transformation is a genetic change in a bacterial or non-animal eukaryotic cell resulting from direct uptake through the cell membrane(s) from its environment and subsequent incorporation of exogenous genetic material (nucleic acid molecules). Transformation can be carried out by artificial means. To transform, cells or bacteria must be in a state of competence, which can manifest as a time-limited response to environmental conditions such as depletion and cell density.

In addition, the present invention provides a host cell transformed or transfi

- 35 042775 cited by a polynucleotide/nucleic acid molecule or a vector according to the invention.

As used herein, the terms host cell or recipient cell are intended to include any single cell or cell culture that can be or is a recipient of vectors, exogenous nucleic acid molecules, and polynucleotides encoding an antibody construct of the invention; and/or recipients of the antibody construct itself. The introduction of the appropriate material into the cell is carried out by transformation, transfection and the like. The term host cell is also intended to include progeny or potential progeny of a single cell. Since certain changes may occur in subsequent generations due to natural, accidental or intentional mutation, or due to environmental influences, such progeny may not actually be completely identical (in morphology or genomic or complete DNA complement) to the parent cell, but all still included in the scope of the term used in this document. Suitable host cells include prokaryotic or eukaryotic cells, and include, but are not limited to, bacteria, yeast cells, fungal cells, plant cells, and animal cells such as insect cells and mammalian cells, e.g., mice, rats, macaques, or humans. .

The antibody construct of the invention can be produced in bacteria. After expression, the antibody construct of the invention is isolated from the E. coli cell paste in a soluble fraction and can be purified by, for example, affinity chromatography and/or size exclusion. The final purification can be performed in a manner analogous to the purification of an antibody expressed, for example, in CHO cells.

In addition to prokaryotes, eukaryotic microorganisms such as filamentous fungi or yeasts are suitable hosts for cloning or expression of antibody constructs of the invention. Saccharomyces cerevisiae, or baker's yeast, is the most widely used host microorganism among lower eukaryotes. However, a number of other genera, species, and strains are generally available and available herein, such as Schizosaccharomyces pombe, Kluyveromyces hosts such as K. lactis, K. fragilis (ATCC 12424), K. bulgaricus (ATCC 16045), K. wickeramii (ATCC 24178), K. waltii (ATCC 56500), K. drosophilarum (ATCC 36906), K. thermotolerans, and K. marxianus; yarrowia (EP 402 226); Pichia pastoris (EP 183 070); Candida Trichoderma reesia (EP 244 234); Neurospora crassa; species of the genus Schwanniomyces such as Schwanniomyces occidentalis and filamentous fungi such as Neurospora, Penicillium, Tolypocladium, and hosts of the genus Aspergillus such as A. nidulans and A. niger.

Suitable host cells for expression of the glycosylated antibody construct of the invention are derived from multicellular organisms. Examples of invertebrate cells include plant and insect cells. Numerous baculovirus strains and variants and corresponding host cells have been identified in susceptible insects such as Spodoptera frugiperda (Foliar armyworm), Aedes aegypti (Yellow fever mosquito), Aedes albopictus (Asian tiger mosquito), Drosophila melanogaster (Drosophila fruit fly) and Bombyx mori (Silkworm). A number of viral strains are available for transfection, for example Autographa californica NPV variant L-1 and Bombyx mori NPV strain Bm-5; such viruses can be used as the virus of the present invention, in particular for transfection of Spodoptera frugiperda cells.

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

However, the greatest interest is shown in relation to vertebrate cells, and the reproduction of vertebrate cells in culture (tissue culture) has become a daily procedure. Examples of suitable mammalian cell lines are CV1 monkey kidney cell line transformed with SV40 (COS-7, ATCC CRL 1651); human kidney embryonic line (293 or 293 cells subcloned for growth in suspension culture as described in Graham et al., J. Gen Virol. 36:59 (1977)); newborn hamster kidney cells (BHK, ATCC CCL 10); Chinese hamster ovary cells/-DHFR (CHO, Urlaub et al., Proc. Natl. Acad. Sci. USA 77: 4216 (1980)); mouse Sertoli cells (TM4, Mather, Biol. Reprod. 23: 243-251 (1980)); monkey kidney cells (CV1 ATCC CCL 70); African green monkey kidney cells (VERO-76, ATCC CRL1587); human cervical carcinoma cells (HELA, ATCC CCL 2); dog kidney cells (MDCK, ATCC CCL 34); buffalo rat liver cells (BRL 3A, ATCC CRL 1442); human lung cells (W138, ATCC CCL 75); human liver cells (Hep G2, 1413 8065); mouse mammary tumor cells (MMT 60562, ATCC CCL5 1); TRI cells (Mather et al., Annals N. Y Acad. Sci. (1982) 383: 44-68); MRC 5 cells; FS4 cells and a human hepatoma line (Hep G2).

In yet another embodiment, the invention relates to a method for producing an antibody construct of the invention, said method comprising culturing a host cell of the invention under conditions to express the antibody constructs of the invention, and recovering the produced antibody construct from the culture.

As used herein, the term cultivation refers to maintaining, diffe

- 36 042775 renirovanie, growth, proliferation and/or propagation of cells in vitro under suitable conditions on the environment. The term expression includes any step involved in the production of an antibody construct of the invention, including, but not limited to, transcription, post-transcriptional modification, translation, post-translational modification, and secretion.

When using recombinant techniques, the antibody construct can be produced intracellularly, in the periplasmic space, or directly secreted into the medium. When using recombinant techniques, the antibody construct can be produced intracellularly, in the periplasmic space, or directly secreted into the medium. Carter et al., Bio/Technology 10: 163-167 (1992) describes a procedure for isolating antibodies secreted into the periplasmic space of E. coli. Briefly, the cell mass is thawed in the presence of sodium acetate (pH 3.5), EDTA, and phenylmethylsulfonyl chloride (PMSF) for approximately 30 minutes. Residual cells can be removed by centrifugation. If the antibody is secreted into the medium, the supernatant from such expression systems is typically first concentrated using commercially available protein concentration filters, such as Amicon or Millipore Pellicon ultrafiltration units. A protease inhibitor such as PMSF may be included to inhibit proteolysis in any of the previous steps, and antibiotics may be added to prevent the growth of incidental contaminants.

An antibody construct of the invention derived from host cells can be isolated or purified using, for example, hydroxyapatite chromatography, gel electrophoresis, dialysis, and affinity chromatography. In addition, depending on the antibody to be purified, other protein purification techniques are available, e.g., ion exchange column fractionation, ethanol precipitation, reverse phase HPLC, heparin-SEPHAROSE™ chromatography, anion or cation exchange resin chromatography (e.g., on a column with polyaspartic acid), chromatofocusing, electrophoresis in SDS-PAGE and precipitation with ammonium sulfate. When an antibody construct of the invention contains a CH3 domain, Bakerbond ABX resin (J.T. Baker, Philipsburg, NJ) is useful for purification.

The preferred purification technique is affinity chromatography. The matrix to which the affinity ligand is attached is most often agarose, but other matrices are also available. Mechanically stable matrices such as controlled pore glass or poly(styrenedivinyl)benzene allow higher flow rates and shorter processing times than with agarose.

In addition, the present invention provides a pharmaceutical composition comprising an antibody construct of the invention or an antibody construct produced in accordance with the method of the invention.

As used herein, the term pharmaceutical composition refers to a composition that is suitable for administration to a patient, preferably a human. A particularly preferred pharmaceutical composition of the invention comprises one or more antibody construct(s) of the invention, preferably in a therapeutically effective amount. Preferably, the pharmaceutical composition further comprises suitable compositions of one or more (pharmaceutically effective) carriers, stabilizers, excipients, diluents, solubilizers, surfactants, emulsifiers, preservatives and/or adjuvants. Acceptable constituents of the composition are preferably non-toxic to recipients at the dosages and concentrations employed. Pharmaceutical compositions of the invention include, but are not limited to, liquid, frozen, and lyophilized compositions.

Compositions of the invention may contain a pharmaceutically acceptable carrier. In general, as used herein, the term pharmaceutically acceptable carrier means any and all aqueous and non-aqueous solutions, sterile solutions, diluents, buffers such as phosphate buffered saline (PBS), water, suspensions, emulsions such as oil/water emulsions, various types of wetting agents, liposomes, dispersion media and coatings that are compatible with pharmaceutical administration, in particular parenteral administration. The use of such media and agents in pharmaceutical compositions is well known in the art, and compositions containing such carriers can be formulated using known conventional methods.

In some embodiments of the invention, pharmaceutical compositions are provided comprising an antibody construct of the invention, and additionally one or more excipients, such as those described in this section and elsewhere in this document. Excipients may be used in this invention in this regard for a wide variety of purposes such as controlling the physical, chemical or biological properties of compositions such as controlling viscosity and/or methods of the invention to improve the performance and/or stabilize such formulations and processes against degradation and spoilage associated, for example, with stresses arising during manufacture, transport, storage, pre-cooking, introduction and after that.

In certain embodiments, the pharmaceutical composition may contain formulation materials to modify, maintain, or maintain, for example, pH, osmolarity, viscosity, clarity, color, isotonicity, odor, sterility, stability, dissolution rate.

- 37 042775 creation or release, absorption or penetration of the composition (see REMINGTON'S PHARMACEUTICAL SCIENCES, 18 Edition, (A.R. Genrmo, ed.), 1990, Mack Publishing Company). In such embodiments, suitable formulation materials may include, but are not limited to:

amino acids such as glycine, alanine, glutamine, asparagine, threonine, proline, 2-phenylalanine, including charged amino acids, preferably lysine, lysine acetate, arginine, glutamate and/or histidine antimicrobial agents such as antibacterial and antifungal agents antioxidants such as ascorbic acid, methionine, sodium sulfite or sodium hydrosulfite;

buffers, buffer systems and buffering agents that are used to maintain the composition at physiological pH or at a slightly lower pH, typically in the pH range of about 5 to about 8 or 9; examples of buffers are borate, bicarbonate, tris-HCl, citrates, phosphates or other organic acids, succinate, phosphate, histidine and acetate; for example, Tris buffer, pH about 7.0-8.5, or acetate buffer, pH about 4.0-5.5;

non-aqueous solvents such as propylene glycol, polyethylene glycol, vegetable oils such as olive oil, and injectable organic esters such as ethyl oleate;

aqueous vehicles, including water, alcoholic/aqueous solutions, emulsions or suspensions, including saline and buffer media;

biodegradable polymers such as polyesters;

bulking agents such as mannitol or glycine;

chelating agents such as ethylenediaminetetraacetic acid (EDTA);

isotonic and absorbent delaying agents;

complexing agents such as caffeine, polyvinylpyrrolidone, beta-cyclodextrin or hydroxypropyl-beta-cyclodextrin) fillers;

monosaccharides; disaccharides; and other carbohydrates (such as glucose, mannose, or dextrins); carbohydrates may be non-reducing sugars, preferably trehalose, sucrose, octasulfate, sorbitol or xylitol;

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

dyes and flavors;

sulfur reducing agents such as glutathione, thioctic acid, sodium thioglycolate, thioglycerol, [alpha]-monothioglycerol and sodium thiosulfate;

diluents;

emulsifiers;

hydrophilic polymers such as polyvinylpyrrolidone;

salt-forming counterions such as sodium;

preservatives such as antimicrobials, antioxidants, chelating agents, inert gases and the like; examples are: benzalkonium chloride, benzoic acid, salicylic acid, thimerosal, phenethyl alcohol, methylparaben, propylparaben, chlorhexidine, sorbic acid or hydrogen peroxide);

metal complexes such as Zn-protein complexes;

solvents and co-solvents (such as glycerin, propylene glycol or polyethylene glycol);

sugars and sugar alcohols such as trehalose, sucrose, octasulfate, mannitol, sorbitol or xylitol, staniose, mannose, sorbose, xylose, ribose, myonisitose, galactose, lactitol, ribitol, myonisitol, galactitol, glycerol, cyclitols (e.g. inositol) glycol ; and polyols of sugars;

suspending agents;

surfactants or wetting agents such as pluronics, PEGs, sorbitan esters, polysorbates such as polysorbate 20, polysorbate, triton, tromethamine, lecithin, cholesterol, tyloxapal; surfactants can be detergents, preferably with a molecular weight > 1.2 kDa and/or polyester, preferably with a molecular weight > 3 kDa; non-limiting examples of preferred detergents are Tween 20, Tween 40, Tween 60, Tween 80 and Tween 85; non-limiting examples for preferred polyesters are PEG 3000, PEG 3350, PEG 4000 and PEG 5000;

stability improving agents such as sucrose or sorbitol;

tonicity enhancing agents such as alkali metal halides, preferably sodium or potassium chloride, mannitol, sorbitol;

parenteral delivery vehicles, including sodium chloride solution, Ringer's dextrose, dextrose and sodium chloride, lactated Ringer's solutions, or fixed oils;

intravenous excipients including fluid replacement solutions

- 38 042775 or nutrients, electrolyte solutions (eg solutions based on Ringer's dextrose).

It will be appreciated by those skilled in the art that the various constituents of a pharmaceutical composition (eg, those listed above) may have different effects, for example, an amino acid may act as a buffer, stabilizer, and/or antioxidant; mannitol can act as a bulking agent and/or tonicity agent; sodium chloride can act as a delivery vehicle and/or tonicity agent; and so on.

It is contemplated that the composition of the invention may contain, in addition to the polypeptide of the invention as defined herein, additional biologically active agents, depending on the intended use of the composition. Such agents may be drugs that act on the gastrointestinal system, drugs that act as cytotoxic drugs, drugs that prevent hyperuricemia, drugs that inhibit immunoreactive reactions (for example, corticosteroids), drugs that modulate the inflammatory response, drugs that acting on the circulatory system and/or agents such as cytokines known in the art. It is also contemplated that the antibody construct of this invention is used in co-therapy, ie in combination with another anti-cancer drug.

In certain embodiments of the invention, the optimal pharmaceutical composition is determined by a person skilled in the art depending on, for example, the route of administration, the delivery format, and the required dosage. See, for example, REMINGTON'S PHARMACEUTICAL SCIENCES, supra. In certain embodiments of the invention, such compositions can affect the physical state, stability, in vivo release rate, and in vivo clearance rate of antibody constructs of the invention. In certain embodiments of the invention, the main medium or carrier in the pharmaceutical composition may be aqueous or non-aqueous in nature. For example, a suitable stock solution or carrier may be water for injection, physiological saline, or artificial cerebrospinal fluid, optionally supplemented with other materials common in parenteral compositions. Neutral buffered saline or saline mixed with serum albumin are additional exemplary media. In certain embodiments of the invention, compositions containing antibody constructs of the invention can be prepared for storage by mixing the selected composition, having the required degree of purity, with optional agents (REMINGTON'S PHARMACEUTICAL SCIENCES, see above) in the form of a lyophilized tablet or aqueous solution. Additionally, in certain embodiments of the invention, the antibody construct of the invention can be obtained as a lyophilisate using appropriate excipients, such as sucrose.

If parenteral administration is envisaged, therapeutic compositions for use in the present invention may be in the form of a pyrogen-free, parenterally acceptable aqueous solution containing the desired antibody construct of the invention in a pharmaceutically acceptable carrier medium. A particularly suitable stock solution for parenteral injection is sterile distilled water to which an antibody construct of the invention has been added as a sterile, isotonic solution with an appropriate preservative. In certain embodiments of the invention, the preparation may include mixing the desired molecule with an agent, such as injectable microspheres, biodegradable particles, polymeric compounds (such as polylactic acid or polyglycolic acid), granules or liposomes, which can provide controlled or prolonged release of the product, which can be delivered via depot injection. In certain embodiments of the invention, you can also use hyaluronic acid, which helps to prolong the period of circulation. In certain embodiments of the invention, implantable drug delivery devices may be used to introduce the desired antibody construct.

The existence of additional pharmaceutical compositions is obvious to those skilled in the art, including sustained or controlled delivery/release dosage forms of the antibody construct of the invention. Methods for preparing various other sustained or controlled delivery agents, such as liposomal vehicles, biodegradable microparticles or porous granules, and depot injections, are also known to those skilled in the art. See, for example, International Patent Application No. PCT/US 93/00829 which describes the controlled release of porous polymeric microparticles for the delivery of pharmaceutical compositions. Sustained release formulations may include semi-permeable polymeric matrices in the form of molded articles such as films or microcapsules. Extended release matrices may include polyesters, hydrogels, polylactides (as described in US Pat. No. 3,773,919 and European Patent Application Publication No. EP 058481), copolymers of L-glutamic acid and gamma-ethyl-E-glutamate (Sidman et al., 1983, Biopolymers 2: 547-556), poly(2hydroxyethyl methacrylate) (Langer et al., 1981, J. Biomed. Mater. Res. 15:167-277 and Langer, 1982, Chem. Tech. 12:98-105), ethylene vinyl acetate (Langer et al., 1981, supra) or poly-O(-)-3-hydroxybutyric acid

- 39 042775 slot (published application for European patent No. EP 133988). Sustained release formulations may also include liposomes, which can be prepared by any of several methods known in the art. See, for example, Eppstein et al., 1985, Proc. Natl. Acad. sci. USA. 82:3688-3692; Published European Patent Application No. EP 036676; EP 088046 and EP 143949.

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

Pharmaceutical compositions used for in vivo administration are generally provided as sterile preparations. Sterilization can be accomplished by filtration through sterile filtration membranes. If the composition is lyophilized, sterilization by this method can be carried out both before and after lyophilization and reconstitution. Compositions for parenteral administration may be stored in lyophilized form or in solution. Parenteral compositions are generally placed in a container with a sterile inlet, for example, an intravenous solution bag or vial with a stopper pierced by a hypodermic injection needle.

Another aspect of the invention includes a self-buffered antibody construct of the invention which can be used as pharmaceutical compositions as described in International Patent Application WO 06138181A2 (PCT/US 2006/022599). A variety of exposures are available for protein stabilization and materials, and preparation methods useful in this regard, such as Arakawa et al., Solvent interactions in pharmaceutical formulations, Pharm Res. 8(3): 285-91 (1991); Kendrick et al., Physical stabilization of proteins in aqueous solution in: RATIONAL DESIGN OF STABLE PROTEIN FORMULATIONS: THEORY AND PRACTICE, Carpenter and Manning, eds. Pharmaceutical Biotechnology. 13: 61-84 (2002), and Randolph et al., Surfactant-protein interactions, Pharm Biotechnol. 13: 159-75 (2002), see especially components and methods for their preparation for self-buffered protein compositions in accordance with the present invention, especially protein pharmaceutical products and processes for veterinary and/or medical purposes.

Salts can be used in accordance with some embodiments of the invention, for example, to adjust the ionic strength and/or isotonicity of the composition and/or to improve the solubility and/or physical stability of the protein or other ingredient of the composition in accordance with the invention. As is known, ions can stabilize the natural state of proteins by binding to charged residues on the surface of the protein and by screening charged and polar groups in the protein and reducing the strength of their electrostatic interactions, attractive and repulsive interactions. The ions can also stabilize the denatured state of the protein by binding, in particular, to the denatured peptide bonds (-CONH) of the protein. In addition, ionic interaction with charged and polar groups in a protein can also reduce intermolecular electrostatic interactions and thereby prevent or reduce protein aggregation and insolubility.

Ionic species vary considerably in their effect on proteins. A number of categorical rankings of ions and their effect on proteins have been developed that can be used in the development of pharmaceutical compositions in accordance with the invention. One example is the Hofmeister series, which ranks ionic and polar non-ionic solutions, influencing the conformational stability of proteins in solution. Stabilizing solutes are called cosmotropic. Destabilizing solutes are called chaotropic. Cosmotropes are typically used at high concentrations (eg > 1 mol ammonium sulfate) to precipitate proteins from solution (salting out). Chaotropes are commonly used to reduce and/or solubilize proteins (salting). The relative effectiveness of ions in salting out and salinity determines their position in the Hofmeister series.

Free amino acids can be used in the construction of antibodies according to the invention in accordance with various embodiments of the invention as excipients, stabilizers and antioxidants, as well as other standard applications. To stabilize the proteins in the preparation, lysine, proline, serine and alanine can be used. Glycine is used during lyophilization to ensure the correct structure and properties of the tablet. Arginine can be used to inhibit protein aggregation in both liquid and lyophilized formulations. Methionine is used as an antioxidant. Polyols include sugars such as mannitol, sucrose and sorbitol, polyhydric alcohols such as, for example, glycerol and propylene glycol, and, for the purposes of this invention, polyethylene glycol (PEG) and related substances.

The polyols are cosmotropic. They are used as stabilizing agents in both liquid and lyophilized formulations to protect proteins from physical and chemical degradation processes. Polyols are also used to adjust the tonicity of drugs. Among the polyols suitable in certain embodiments of the invention, there is mannitol, usually used

- 40 042775 used to ensure the structural stability of the sediment in lyophilized formulations. It ensures the stability of the structure to precipitation. It is generally used in conjunction with a lyoprotectant such as sucrose. Sorbitol and sucrose are among the preferred tonicity adjusting agents and as stabilizers to protect against freeze/thaw stress during shipping or bulk preparation during the manufacturing process. Reducing sugars (containing free aldehyde or ketone groups), such as glucose and lactose, can glycate surface lysines and arginine residues. Therefore, they are generally not among the preferred polyols for use in accordance with the invention. In addition, sugars which form reactive species such as sucrose, which hydrolyze to fructose and glucose under acidic conditions and hence cause glycation, are also not among the preferred polyols of the invention in this respect. PEG is used to stabilize proteins and as a cryoprotectant, and in this regard it can be used in this invention.

Embodiments of the antibody construct of the invention also include surfactants. Protein molecules can be subject to adsorption on surfaces and denaturation and subsequent aggregation at air-liquid, solid-liquid, and liquid-liquid interfaces. These effects are generally inversely proportional to protein concentration. These detrimental interactions are generally inversely related to protein concentration and are typically exacerbated by physical shaking, such as occurs during product handling and shipping. Surfactants are typically used to prevent, minimize or reduce such surface adsorption. Suitable surfactants in the invention in this regard include polysorbate 20, polysorbate 80, other fatty acid esters of sorbitan polyethoxylates, and poloxamer 188. Also, surfactants have traditionally been used to regulate the conformational stability of proteins. The use of surfactants in this regard is specific to the protein, since any given surfactant will generally stabilize some proteins and destabilize others.

Polysorbates are susceptible to oxidative degradation and are often said to contain sufficient peroxides to cause oxidation of protein side chains, especially methionine. Therefore, polysorbates should be used with caution and, when used, should be used at their lowest effective concentration. In this regard, polysorbates represent the general rule that excipients should be used at their lowest effective concentrations.

Embodiments of the antibody construct of the invention further include one or more antioxidants. To some extent, harmful protein oxidation can be prevented in pharmaceutical compositions by maintaining proper ambient oxygen levels and temperatures and avoiding exposure to light. Antioxidants can also be used to prevent oxidative degradation of proteins. Suitable antioxidants in this respect include reducing agents, oxygen/free radical scavengers and chelating agents. Antioxidants for use in therapeutic protein compositions according to the invention are preferably water soluble and retain their activity throughout the shelf life of the product. EDTA is the preferred antioxidant according to the invention in this regard. Antioxidants can damage proteins. For example, reducing agents such as glutathione in particular can disrupt intramolecular disulfide bonds. Thus, antioxidants for use in the invention are selected, among other things, to eliminate or sufficiently reduce the likelihood that they themselves will damage the proteins in the composition.

The preparations according to the invention may contain metal ions which are protein coexductors and which are necessary for the formation of protein coordination complexes, such as zinc, which is necessary for the formation of certain insulin suspensions. Metal ions can also inhibit some of the processes in which protein degradation occurs. However, metal ions also catalyze physical and chemical processes that degrade proteins. Magnesium ions (10120 mm) can be used to suppress the isomerization of aspartic acid to isoaspartic acid. Ca ⁺ ions (up to 100 mM) can increase the stability of human deoxyribonuclease. Mg ⁺ , Mn ⁺ and Zn ⁺ , however, can destabilize human deoxyribonuclease (rhDNase). By analogy, Ca ⁺² and Sr ⁺² can stabilize factor VIII, it can be destabilized with Mg ⁺² , Mn ⁺² and Zn ⁺² , Cu ⁺² and Fe ⁺² ions, and its aggregation can be increased with Al ⁺³ .

Embodiments of the antibody construct of the invention further include one or more preservatives. Preservatives are needed in the development of multi-dose parenteral products for which more than one set from the same container is provided. Their main function is to inhibit the growth of microorganisms and guarantee the sterility of the product during the expiration date or term of use of the medicinal product. Traditionally used preservatives include benzyl alcohol, phenol and m-cresol. Although preservatives have a long history of use with small molecule parenterals, the development of protein formulations

- 41 042775 holding preservatives can be a challenge. Preservatives almost always have a destabilizing effect (aggregation) on proteins, and this has become a major factor limiting their use in multi-dose protein formulations. To date, most protein formulations have been developed for single use only. However, when multi-dose formulations are possible, they have the added benefit of providing patient convenience and increased competitiveness. A good example is human growth hormone (hGH), where the development of conserved formulations has led to the commercialization of more convenient, reusable pen-format injection devices. There are currently at least four such pen-sized devices available on the market containing stored hGH formulations. Norditropin (Liquid, Novo Nordisk), Nutropin AQ (Liquid, Genentech) & Genotropin (Lyophilized - Dual Chamber Cartridge, Pharmacia & Upjohn) contain phenol while Somatrope (Eli Lilly) is formulated with m-cresol. When developing and creating canned dosage forms, several aspects must be considered. The effective concentration of the preservative in the drug product must be optimized. This calls for testing this preservative in dosage form with concentration ranges that confer antimicrobial efficacy without compromising protein stability.

As would be expected, the development of liquid formulations containing preservatives is more complex than lyophilized formulations. Freeze-dried products may be freeze-dried without a preservative and reconstituted with a diluent-containing preservative at the time of use. This reduces the time the preservative is in contact with the protein, greatly minimizing the associated stability risks. When using liquid formulations, preservative effectiveness and stability must be maintained throughout the shelf life of the product (18 to 24 months). It is important to note that the effectiveness of the preservative must be demonstrated in the final formulation containing the active drug and all excipient components.

The antibody constructs described herein can also be formulated as immunoliposomes. A liposome is a small vesicle composed of various types of lipids, phospholipids and/or surfactants that are suitable for drug delivery to a mammal. The components of a liposome are typically organized in a bilayer composition, similar to the distribution of lipids in biological membranes. Liposomes containing antibody constructs are prepared using methods known in the art, such as those described in Epstein et al., Proc. Natl. Acad. sci. USA 82: 3688 (1985); Hwang et al., Proc. Natl Acad. sci. USA 77: 4030 (1980); US patent No. 4485045 and 4544545 and WO 97/38731. Liposomes with extended circulation time are described in US Pat. No. 5,013,556. Particularly suitable liposomes can be prepared using a reverse phase evaporation process using a lipid composition containing phosphatidylcholine, cholesterol, and PEG-derived phosphatidylethanolamine (PEG-PE). Liposomes are extruded through filters with a defined pore size to obtain liposomes having a target diameter. Fab' fragments of an antibody construct of the invention can be conjugated to liposomes as described in Martin et al. J Biol. Chem. 257: 286-288 (1982) via a disulfide exchange reaction. The chemotherapeutic agent is optionally contained within a liposome. See Gabizon et al., J. National Cancer Inst. 81 (19) 1484 (1989).

Once prepared, the pharmaceutical composition may be stored in a sterile vial as a solution, suspension, gel, emulsion, solid, crystal, or dehydrated lyophilized powder. Such preparations can be stored both in ready-to-use form and in a form (eg, lyophilized) that is reconstituted prior to administration.

The biological activity of a pharmaceutical composition as defined herein can be determined, for example, by cytotoxicity assay as described in the following examples, in WO 99/54440 or in 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, for example, standardized NCI response criteria. Efficacy or in vivo efficacy of therapy using a pharmaceutical composition of the invention refers to the effectiveness of the composition for its intended purpose, ie the ability of the composition to produce its desired effect, i.e. depletion of pathological cells, eg tumor cells. In vivo efficacy can be monitored by established standard methods for appropriate disease subjects including, but not limited to, white blood cell count, differentials, fluorescent activated cell sorting, bone marrow aspiration. In addition, various clinical disease chemistry parameters and other established standard methods may be used. In addition, computed tomography, x-ray, nuclear magnetic resonance imaging (eg, to evaluate results based on the National Cancer Institute [Cheson BD, Horning SJ, Coiffier B, Shipp MA, Fisher RI, Connors JM, Lister TA, Vose J, GrilloLopez A , Hagenbeek A, Cabanillas F, Klippensten D, Hiddemann W, Castellino R, Harris NL, Armitage JO, Carter W, Hoppe R, Canellos GP Report of an international workshop to standardize response criteria for nonHodgkin's lymphomas NCI Sponsored International Working Group. J Clin Oncol. 1999 Apr; 17(4):1244]), positron emission tomography, leukocyte count, differentials, flu

- 42 042775 orescently activated cell sorting, bone marrow aspiration, biopsy/histology of lymph nodes and various clinical chemistry parameters of lymphoma (eg lactate dehydrogenase) and other established standard methods.

Another important goal in drug development, such as the pharmaceutical composition of the invention, is the predictable modulation of pharmacokinetic properties. To this end, it is possible to establish a pharmacokinetic profile of a drug candidate, that is, a profile of pharmacokinetic parameters that affect the ability of a particular drug to treat a given condition. Pharmacokinetic parameters of a drug that affect the ability of a drug to treat a particular disease include, but are not limited to: half-life, volume of distribution, first-pass metabolism, and serum binding. Each of the above parameters can affect the effectiveness of this drug.

Half-life refers to the time at which 50% of an administered drug is eliminated by biological processes such as metabolism, excretion, etc. By first pass hepatic metabolism is meant the propensity of a drug to be metabolized upon first contact with the liver, i.e. during its first passage through the liver. Volume of distribution refers to the degree of retention of a drug in all compartments of the body, eg, intracellular and extracellular spaces, tissues and organs, etc., and the distribution of the drug in these compartments. The degree of binding in blood serum means the propensity of the drug to interact with blood serum proteins and their binding to blood serum proteins, such as albumin, resulting in a decrease or loss of the biological activity of the drug.

Pharmacokinetic parameters also include bioavailability, delay time (Tlag), Tmax, rate of absorption, greater appearance and/or Cmax for a given amount of drug administered. Bioavailability refers to the amount of drug in the blood compartment. Lag time means the time delay between administration of a drug and its detection and measurability in blood or plasma. Tmax is the time after which the maximum blood concentration of the drug is reached, and Cmax is the maximum blood concentration obtained with the drug. The time to reach the concentration of the drug in the blood or tissue, which is required for its biological effect, depends on all parameters. Pharmacokinetic parameters of cross-species specific bispecific antibody constructs that can be determined in preclinical animal studies in non-chimpanzee primates as described above are also set forth, for example, in Schlereth et al. (Cancer Immunol. Immunother. 20 (2005), 1-12).

In one embodiment, the invention provides an antibody construct of the invention, or an antibody construct prepared in accordance with the method of the invention, for use in the prevention, treatment, or amelioration of cancer.

The formulations described herein are useful as pharmaceutical compositions for the treatment, amelioration and/or prevention of a medical condition as described herein in a patient in need thereof. The term treatment refers to both therapeutic intervention and prophylactic or preventive measures. Treatment involves the use or administration of a drug into an organism, isolated tissue, or cell in a patient who has a disease/disorder, a symptom of a disease/disorder, or a predisposition to a disease/disorder, for the purpose of curing, treating, partially relieving symptoms, alleviating, modifying, eliminating, improving condition, alleviate or influence a disease, a symptom of a disease, or a predisposition to a disease.

As used herein, the term amelioration refers to any improvement in the disease state of a patient having one of the types of (metastatic) tumors or cancers as defined herein by administering an antibody construct according to the invention to a subject in need thereof. Such an improvement can also be considered as slowing down or stopping the progression of a (metastatic) tumor or cancer in a patient. As used herein, the term prophylaxis means preventing the occurrence or reappearance of a patient having one of the types of (metastatic) tumors or cancers as defined herein by administering an antibody construct according to the invention to a subject in need thereof.

The term disease refers to any condition that could be corrected by treatment with an antibody construct or a pharmaceutical composition as described herein. This term encompasses chronic and acute disorders or diseases, including those pathological conditions that increase the susceptibility of a mammal to the disease in question.

A neoplasm is an abnormal growth of tissue, usually but not always forming a mass. When a mass also forms, it is usually referred to as a tumor. Neoplasms or tumors

- 43 042775 may be benign, potentially malignant (precancerous), or malignant. Malignant neoplasms are commonly referred to as cancer. They usually invade and destroy surrounding tissue and may form metastases, i.e. they spread to other parts, tissues, or organs of the body. Therefore, the term metastatic cancer covers metastases to other tissues or organs, as opposed to the original tumor. Lymphomas and leukemias are lymphoid neoplasms. For the purposes of this invention, they are also covered by the terms tumor or cancer.

In a preferred embodiment, the invention provides an antibody construct of the invention, or an antibody construct prepared according to the method of the invention, for use in the prevention, treatment, or amelioration of cancer, the cancer being selected from the group consisting of lung carcinoma, head and neck carcinoma , primary or secondary CNS tumor, primary or secondary brain tumor, primary CNS lymphoma, spinal axis tumors, brainstem glioma, glioblastoma, pituitary adenoma, adrenocortical cancer, esophageal carcinoma, colon cancer, breast cancer, ovarian cancer, NSCLC ( non-small cell lung cancer), SCLC (small cell lung cancer), endometrial cancer, cervical cancer, uterine cancer, transitional cell carcinoma, bone cancer, pancreatic cancer, skin cancer, cutaneous or intraocular melanoma, liver cancer, bile duct cancer, gallbladder cancer , kidney cancer, rectal cancer, anal cancer, gastric cancer, gastrointestinal (gastric, colorectal and duodenal) cancer, small intestine cancer, bile duct cancer, urethra cancer, renal cell carcinoma, endometrial carcinoma, thyroid cancer, testicular cancer, squamous cell skin cancer, melanoma, gastric cancer, prostate cancer, bladder cancer, osteosarcoma, mesothelioma, Hodgkin's disease, non-Hodgkin's lymphoma, chronic or acute leukemia, chronic myeloid leukemia, lymphocytic lymphomas, multiple myeloma, fibrosarcoma, neuroblastoma, retinoblastoma and soft tissue sarcomas, and metastatic cancer derived from any of the above. The cancer (metastatic) is preferably P-cadherin positive or P-cadherin expressing cancer.

In another preferred embodiment, the invention provides an antibody construct of the invention, or an antibody construct prepared according to the method of the invention, for use in the prevention, treatment, or amelioration of a cancer, the cancer being (metastatic) squamous cell carcinoma.

The invention also provides a method of treating or improving a (metastatic) tumor or cancer, comprising the step of administering to a subject in need thereof an antibody construct of the invention or an antibody construct prepared according to the method of the invention.

In a preferred embodiment, the invention provides a method of treating or improving a cancer, wherein the cancer is selected from the group consisting of lung carcinoma, head and neck carcinoma, primary or secondary CNS tumor, primary or secondary brain tumor, primary CNS lymphoma, axis tumors. spine, brainstem glioma, glioblastoma, pituitary adenoma, adrenocortical cancer, esophageal carcinoma, colon cancer, breast cancer, ovarian cancer, NSCLC (non-small cell lung cancer), SCLC (small cell lung cancer), endometrial cancer, cervical cancer, cancer uterus, transitional cell carcinoma, bone cancer, pancreatic cancer, skin cancer, skin or intraocular melanoma, liver cancer, bile duct cancer, gallbladder cancer, kidney cancer, rectal cancer, anal cancer, gastric cancer, gastrointestinal (gastric, colorectal and duodenal) cancer, small intestine cancer, bile duct cancer, urethral cancer, renal cell carcinoma, endometrial carcinoma, thyroid cancer, testicular cancer, squamous cell skin cancer, melanoma, gastric cancer, prostate cancer, urinary cancer bladder, osteosarcoma, mesothelioma, Hodgkin's disease, non-Hodgkin's lymphoma, chronic or acute leukemia, chronic myeloid leukemia, lymphocytic lymphomas, multiple myeloma, fibrosarcoma, neuroblastoma, retinoblastoma and soft tissue sarcoma and metastatic cancer derived from any of the above, comprising the step of administering to a subject in need thereof, an antibody construct of the invention or an antibody construct prepared in accordance with the method of the invention.

In yet another preferred embodiment, the present invention provides a method of treating or improving a tumor or cancer, or a metastatic tumor or cancer, wherein the cancer is (metastatic) squamous cell carcinoma. The terms subject in need or in need of treatment include those who already have the disorder as well as those in whom the disorder is to be prevented. A subject in need or diseased includes humans and other mammalian subjects who are receiving either prophylactic or therapeutic treatment.

An antibody design in accordance with the present invention is generally designed for specific routes and routes of administration, for specific dosages and frequencies of administration, for specific treatment of specific diseases, with ranges of bioavailability and persistence, among others. The materials of the composition are preferably formulated in concentrations suitable for the site of administration.

- 44 042775

Thus, preparations and compositions can be designed in accordance with the invention for delivery by any suitable route of administration. In the context of this invention, routes of administration include, but are not limited to, topical application (such as dermal, inhalation, nasal, ophthalmic, auricular/auditory, vaginal, mucosal);

enteral routes (such as oral, gastrointestinal, sublingual, sublabial, buccal, rectal) and parenteral routes (such as intravenous, intraarterial, intraosseous, intramuscular, intracerebral, intracranial, epidural, spinal, subcutaneous, intraperitoneal, extraamniotic, intraarticular, intracardiac , intradermal, intralesional, intrauterine, intravesical, intravitreal, transdermal, intranasal, transmucosal, intrasynovial, intraluminal).

Pharmaceutical compositions and antibody constructs of this invention are particularly useful for parenteral administration, such as subcutaneous or intravenous delivery, for example, by injection, such as a bolus injection, or by infusion, such as continuous infusion. Pharmaceutical compositions can be administered using a medical device. Examples of medical devices for administering pharmaceutical compositions are described in US Pat. Nos. 4,475,196; 4439196; 4447224; 4447233; 4486194; 4487603; 4596556; 4790824; 4941880; 5064413; 5312335; 5312335; 5383851 and 5399163.

In particular, the present invention provides for the continuous administration of a suitable composition. As a non-limiting example, continuous or substantially continuous, ie continuous, administration can be accomplished with a small pump system carried by the patient to measure the influx of the therapeutic agent into the patient's body. A pharmaceutical composition containing an antibody construct of the invention may be administered using said pumping systems. Such pumping systems are generally known in the art and typically rely on periodic replacement of cartridges containing the therapeutic agent to be administered. When replacing a cartridge in such a pumping system, a temporary interruption of the otherwise uninterrupted flow of the therapeutic agent to the patient may occur. In such a case, the administration phase before cartridge replacement and the administration phase after cartridge replacement will still be considered in the sense of the pharmaceutical agents and methods of the invention together constitute one continuous administration of such a therapeutic agent.

Sustained or continuous administration of the antibody constructs of the invention may be intravenous or subcutaneous via a fluid delivery device or a small pumping system containing a fluid delivery mechanism to remove fluid from a reservoir and an actuator to actuate the delivery mechanism. Subcutaneous pump systems may include a needle or cannula for penetrating a patient's skin and delivering a single composition into the patient's body.

Said pumping systems can be directly fixed or affixed to the patient's skin independently of a vein, artery or blood vessel, thereby providing direct contact between the pumping system and the patient's skin. The pump system can be attached to the patient's skin from 24 hours to several days. The pumping system can be small in size with a reservoir for small volumes. As a non-limiting example, the volume of the reservoir for a suitable pharmaceutical composition to be administered may be from 0.1 to 50 ml.

Long-term administration can also be transdermal by patching the skin and changing at intervals. One of skill in the art is aware of drug delivery patch systems suitable for this purpose. It should be noted that transdermal administration is particularly suitable for continuous administration, since the exchange of the first expended patch can be successfully carried out simultaneously with the placement of a new second patch, for example, on the surface of the skin immediately adjacent to the first expended patch and immediately before the removal of the first expended patch. There are no problems with flow interruption or battery failure.

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

Compositions of this invention may be administered to a subject at a suitable dose, which can be determined, for example, by dose escalation studies when administering increased doses of an antibody construct of the invention exhibiting cross-species specificity as described herein to primates other than chimpanzees, for example, macaques. As indicated above, an antibody construct of the invention exhibiting cross-species specificity as described herein can advantageously be used in an identical form in preclinical studies in

- 45 042775 mats other than chimpanzees and as a drug in humans. 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 anthropometric data, body surface area, age, specific compound to be administered, sex, time and route of administration, general health, and other drugs being administered. simultaneously.

The term effective dose or effective dosage is defined as an amount sufficient to achieve or at least partially achieve the desired effect. The term therapeutically effective dose is defined as an amount sufficient to treat or at least partially arrest a disease and its complications in a patient already suffering from the disease. Amounts or doses effective for this use will depend on the condition being treated (indication), the antibody design being set, the therapeutic context and goals, the severity of the disease, prior therapy, the patient's medical history and response to the therapeutic agent, route of administration, size (weight body area or organ size) and/or the condition (age and general health) of the patient, as well as the general condition of the patient's own immune system. The correct dose may be adjusted according to the judgment of the attending physician so that it may be administered to the patient one or more times over several administrations and to achieve the optimum therapeutic effect.

A typical dosage may be in the range of about 0.1 μg/kg to about 30 mg/kg or more, depending on the above factors. In specific embodiments, the dosage may be in the range of 1.0 µg/kg to about 20 mg/kg, in some cases 10 µg/kg to about 10 mg/kg, or 100 µg/kg to about 5 mg/kg.

A therapeutically effective amount of an antibody construct of the invention preferably results in a reduction in the severity of symptoms of the disease, an increase in the frequency or duration of periods without symptoms of the disease, or the prevention of deterioration or disability due to the disease. For the treatment of CDH3-expressing tumors, a therapeutically effective amount of an antibody construct of the invention, e.g., an anti-CDH3/anti-CD3 antibody, preferably inhibits cell or tumor growth by at least about 20%, at least about 40%, at least at least about 50%, at least about 60%, at least about 70%, at least about 80%, or at least about 90% compared to untreated patients. The ability of a compound to inhibit tumor growth can be evaluated in an animal model predictive of efficacy in human tumors.

The pharmaceutical composition can be administered as a single therapeutic agent or in combination with additional therapies, such as anti-cancer therapies as needed, for example, other protein and non-protein drugs. These drugs can be administered simultaneously with a composition containing an antibody construct of the invention as defined herein, or separately before or after administration of said antibody construct at timely defined intervals and doses.

As used herein, the term effective and non-toxic dose refers to the tolerable dose of an antibody construct of the invention that is high enough to cause depletion of pathological cells, tumor eradication, tumor shrinkage, or disease stabilization without or substantially without significant toxic effects. Such effective and non-toxic doses can be determined, for example, using dose escalation studies described in the art, and should be below the dose that causes serious adverse side effects (dose limiting toxicity - DLT).

As used herein, the term toxicity refers to the toxic effect of a drug resulting in adverse events or severe adverse events. These side effects may relate to the lack of tolerability of the drug in general and / or the lack of local tolerance after administration. Toxicity may also include teratogenic or carcinogenic effects caused by the drug.

The term safety, in vivo safety, or tolerability, as used herein, defines the administration of a drug without causing serious side effects immediately after administration (local tolerance) and over a longer period of drug use. Safety, in vivo safety or tolerability can be assessed, for example, at regular intervals during the treatment and follow-up period. Measurements include clinical assessment, such as organ manifestations and screening for laboratory abnormalities. Clinical evaluation can be performed and abnormal results recorded/coded according to NCI-CTC and/or MedDRA standards. Organ manifestations may include criteria such as allergy/immunology, blood/bone marrow, cardiac arrhythmia, coagulation, and the like, as specified, for example, in the Common Criteria for Adverse Events v3.0 (CTCAE). Laboratory parameters that may be tested include, for example, hematology, clinical chemistry, coagulation profile, and urinalysis and examination of other body fluids such as serum, plasma, lymphoid or spinal fluid, cerebrospinal fluid, and the like. So

- 46 042775 way, safety can be assessed, for example, by physical examination, imaging methods (i.e. ultrasound, x-ray, computed tomography, magnetic resonance imaging (MRI), other measurements using technical devices (e.g. electrocardiogram), vital signs by measuring laboratory parameters and recording adverse events For example, adverse events in non-chimpanzee primates in the uses and methods of the invention can be investigated by histopathological and/or histochemical methods.

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

In yet another embodiment, the invention provides a kit comprising an antibody construct of the invention, an antibody construct prepared according to a method of the invention, a vector of the invention, and/or a host cell of the invention. In the context of this invention, the term kit means two or more components, one of which corresponds to an antibody construct, pharmaceutical composition, vector, or host cell of the invention, packaged together in a container, reservoir, or the like. Thus, a kit can be described as a set of products and/or tools that are sufficient to achieve a specific purpose and that can be sold as a unit.

The kit may contain one or more reservoirs (such as vials, ampoules, containers, syringes, bottles, bags) of any suitable shape, size and material (preferably waterproof, such as plastic or glass) containing the antibody construct or pharmaceutical composition of this invention in appropriate dosage for administration (see above). The kit may additionally contain instructions for use (for example, in the form of a brochure or instructions for use), means for administering the antibody construct of this invention, such as a syringe, pump, infusor, or the like, a means for reconstituting the antibody construct of the invention, and/or a means to dilute the antibody construct of the invention. The invention also provides single dose kits for administration. The kit of the invention may also contain a first reservoir containing a dry/lyophilized antibody construct and a second reservoir containing an aqueous composition. In some embodiments of the present invention, kits are provided containing single-chamber and multi-chamber pre-filled syringes (eg, liquid syringes and lyophilisate syringes).

Description of graphic materials

Fig. 1: Schematic representation of the five D1-D5 extracellular domains of human CDH3, mouse CDH3, and one exemplary chimeric CDH3 (in this document: mouse D1B). The five domains are again shown below, and further divided into three subdomains each. An interpretation of a typical FACS signal (see epitope clustering in Example 2) is shown at the bottom right.

Fig. 2: Human CDH3 and mouse CDH3 alignment sequence and typical display of different domains: signal peptide, propeptide, D1-D5 extracellular domains, transmembrane domain and cytoplasmic domain. Mouse sequence-based substitutions of the five extracellular domains were introduced into the human CDH3 scaffold, see example 1, and then the chimeric constructs were used for epitope clustering (epitope mapping), see example 2.

Fig. 3: Human and mouse CDH3 and 20 human/mouse CDH3 chimeric constructs (five extracellular domains (ECDs) and three subdomains for each ECD) expressed on the surface of CHO cells as shown by flow cytometry. Expression of human wild-type CDH3, mouse wild-type CDH3, and CDH3 chimeric constructs on CHO cells was tested with a monoclonal mouse anti-human CDH3 IgG1 antibody that is cross-reactive. Bound monoclonal antibody was detected with anti-mouse IgG Fc-PE (1:100, 50 μl; Jackson Immunoresearch no. 115-116-071). D1*): Hu CDH3 D1 mu-CHO.

Fig. 4: Epitope mapping of CDH3 constructs. For examples of binding molecules specific to different epitope clusters/non-cellular subdomains detected by epitope mapping, CDH3 chimeric constructs, see Example 2.

Fig. 4A: D1B binding domains.

Fig. 4B: D2C binding domains.

Fig. 4C: D3A binding domains.

Fig. 5:

FACS binding assay of 5 μg/ml of purified bispecific antibody monomer on indicated cell lines. See also Example 5. Detection of CDH3xCD3 bispecific antibody binding was performed with native mouse antibodies specific for the CD3 binding portion of the bispecific antibody followed by goat anti-mouse Fcy-PE. A negative control was performed with buffer only, followed by detection of antibodies.

Fig. 5A and 5B: CDH3xCD3 bispecific binding domains (FIG. 5A: D2C epitope cluster/extracellular subdomain; FIG. 5B: D3A epitope cluster/extracellular subdomain) were

- 47 042775 were analyzed for their binding to CHO cells transfected with human CDH3, human CD3 on the HPB-all T cell line, CHO cells transfected with macaque CDH3, macaque CD3 on the HSC-F T cell line expressing macaque CD3, CDH3-positive line A431 human carcinoma cells and mouse CDH3-transfected CHO cells (negative control). Binding was found in all cases except for the negative control.

Fig. 5C and 5D: CDH3xCD3 bispecific binding domains (FIG. 5C: D2C epitope cluster/extracellular subdomain; FIG. 5D: D3A epitope cluster/extracellular subdomain) were analyzed for their binding to human CDH3 paralogs CDH1, CDH2, CDH4 and CDH5. Paralog binding was not found. No binding of dhfr ^−/ 7 by CHO cells (negative control) was detected.

Fig. 6: Cytotoxic activity of stimulated human CD8+ T cells against CHO cells transfected with human CDH3 in the presence of CDH3xCD3 bispecific antibodies measured in an 18 hour chromium release assay. Effector cells: stimulated enriched human CD8+ T cells. Target cells: CHO cells transfected with human CDH3. Effector cell:target cell ratio (E:T): 10:1. Antibodies specific for the epitope cluster/extracellular subdomain of D2C (Fig. 6A) and D3A (Fig. 6B)

Fig. 7: Cytotoxic activity of stimulated human CD8+ T cells against CDH3 positive human epidermoid carcinoma cell line A431 in the presence of CDH3xCD3 bispecific antibodies measured in an 18 hour chromium release assay. Effector cells: stimulated enriched human CD8+ T cells. Target cells: human A431 cells. Effector cell:target cell ratio (E:T): 10:1. Antibodies specific for the epitope cluster/extracellular subdomain of D2C (FIG. 7A) and D3A (FIG. 7B).

Fig. 8: Cytotoxic activity of unstimulated human PBMCs against CHO cells transfected with human CDH3 in the presence of CDH3xCD3 bispecific antibodies measured in a 48-hour FACS-based cytotoxicity assay. Effector cells: unstimulated human PBMC (CD14-/CD56-). Target cells: CHO cells transfected with human CDH3. Effector cell:target cell ratio (E:T): 10:1. Antibodies specific for the epitope cassette/extracellular subdomain of D2C (FIG. 8A) and D3A (FIG. 8B).

Fig. 9: Cytotoxic activity of unstimulated human PBMC against CDH3-positive human epidermoid carcinoma cell line A431 in the presence of CDH3xCD3 bispecific antibodies measured in a 48-hour FACS-based cytotoxicity assay. Effector cells: unstimulated human PBMC (CD14-/CD56-). Target cells: human A431 cells. Effector cell:target cell ratio (E:T): 10:1. Antibodies specific for the epitope cluster/extracellular subdomain of D2C (Fig. 9A) and D3A (Fig. 9B)

Fig. 10: Cytotoxic activity of a macaque T cell line against CHO cells transfected with macaque CDH3 in the presence of CDH3xCD3 bispecific antibodies measured in a 48-hour FACS-based cytotoxicity assay. Efferent cells: CD3-positive macaque T cell line LnPx4119. Target cells: CHO cells transfected with macaque CDH3. Effector cell:target cell ratio (E:T): 10:1. Antibodies specific for epitope cassette/extracellular subdomain D2C (Fig. 10A), D3A (Fig. 10B), and D1B (Fig. 10C)

Fig. 11: Antitumor activity of the D2C epitope cassette/extracellular subdomain CDH3xCD3 bispecific antibody (CDH3-13) in a human tumor xenograft model (see Example 14). The antibody dose-dependently prevents the formation of A-431 tumors in the presence of human PBMCs. The high volume of the tumor at the beginning of the measurement (5th day) due to the large volume of the cell mixture administered on day 1. **p<0.01; ***p<0.0001.

Fig. 12: Antitumor activity of the bispecific and extended half-life (HLE) antibody CDH3xCD3 epitope cassette/extracellular subdomain D2C (CDH313) in a human tumor xenograft model (see Example 15). The HLE antibody dose-dependently prevents the formation of human HCT-116 tumors in the presence of human PBMCs. Although in FIG. 12A shows the overall result, FIG. 12B highlights the result obtained for higher antibody concentrations (group 2) in responders (7/10) and non-responders (3/10).

Fig. 13: T cell activation in the absence of target cells with bispecific antibody constructs in the absence (upper panel) and presence (lower panel) of albumin fusion at the C-terminus of the construct.

Examples

The following examples illustrate the invention. These examples are not to be construed as limiting the scope of the present invention. This invention is limited only by the claims.

Example 1 Generation of CHO Cells Expressing Wild Type and Chimeric CDH3

To construct chimeric molecules used for epitope mapping, the sequence of the respective five extracellular domains Dom1-Dom5 (or D1-D5) and their subdomains (A,

- 48 042775

B and C) Human CDH3 was replaced with the corresponding mouse sequence. The following 20 molecules were obtained; see also FIG. 1 and 2:

Hu CDH3/Doml mu (ak 108-215)

Nor CDH3/DomlAmu (ak 108-143)

Nor CDH3/DomlB mu (ak 144-179)

Nor CDH3/DomlC mu (ak 180-215)

Nor CDH3/Dom2 mu (ak 216-327)

Nor CDH3/Dom2Amu (ak 216-252)

Nor CDH3/Dom2B mu (ak 253-290)

Nor CDH3/Dom2C mu (ak 291-327)

Nor CDH3/Dom3 mu (ak 328-440)

Nor CDH3/Dom3Amu (ak 328-363)

Nor CDH3/Dom3B mu (ak 364-403)

Nor CDH3/Dom3C mu (ak 404-440)

Neither CDH3 / Dom4 mu (aa 441-546)

Nor CDH3/Dom4Amu (ak 441-474)

Nor CDH3/Dom4B mu (ak 475-511)

Nor CDH3/Dom4C mu (ak 512-546)

Nor CDH3/Dom5 mu (ak 547-650)

Nor CDH3/Dom5Amu (ak 547-581)

Nor CDH3/Dom5B mu (ak 582-616)

Nor CDH3/Dom5C mu (ak 617-650)

SEQ W NO: 13

SEQ ID NO: 14

SEQ ID NO: 15

SEQ ID NO: 16

SEQ ID NO: 17

SEQ ID NO: 18

SEQ ID NO: 19

SEQ ID NO: 20

SEQ ID NO: 21

SEQ ID NO: 22

SEQ ID NO: 23

SEQ ID NO: 24

SEQ ID NO: 25

SEQ ID NO: 26

SEQ ID NO: 27

SEQ ID NO: 28

SEQ ID NO: 29

SEQ ID NO: 30

SEQ ID NO: 31

SEQ ID NO: 32

The above list shows the positions of the various domains (Dom1-Dom5) as well as the corresponding A-C subdomains in the human CDH3 amino acid sequence as shown in SEQ ID NO: 1. For example, the D1A subdomain is located at amino acid positions 108-143 of SEQ ID NO : 1. The same applies to all other domains listed above.

For expression in CHO cells, the coding sequence of the chimeric extracellular domains described above was followed in-frame by an artificial Ser/Glylinker coding sequence followed by a domain derived from the transmembrane/intracellular domain of human EpCAM (amino acids 266-314 of the sequence published with GenBank registration number NM_002354) . All chimeric constructs consisted of an N-terminal signal sequence (signal peptide) and a propeptide.

To generate CHO cells expressing human CDH3, mouse cynomolgus (cyno) CDH3, and human/mouse chimeric CDH3, the respective coding sequences for human CDH3 (SEQ ID NO: 2, see also GeneBank accession number NM_001793), macaque CDH3 (SEQ ID NO: 6), mouse CDH3 (SEQ ID NO: 10, see also GeneBank accession number NM_001037809) and 20 human/mouse CDH3 chimeras (see above) were cloned into a plasmid designated pEF-DHFR (pEF-DHFR described in Raum et al. Cancer Immunol Immunother 50 (2001) 141-150). The macaque CDH3 coding sequence was obtained by standard cloning using the macaque spleen cDNA library (BioChain) and human sequence-specific oligonucleotides (5' GGCCCGCCGTCGCGGCAGC 3'; 5' CTCCTTCTCCAGGTTTGCTGGC 3'; 5' AACTGAGA

CCCTTGGAGATTGC 3'; 5' TAGTCGTCCTCCCCGCCACC 3'; 5’ GGAGGGTGGGACAAACA

CAGG3'; 5' ACGTTGAAGTGACCAACGAGGC 3') hybridized in the untranslated region or region of the conserved sequence of the human CDH3 mRNA transcript (NM_001793). Sequence analysis showed amino acid sequence similarity of the major extracellular domain to GenHank CDH3 sequences (JU473826, JU473827). All cloning 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)). For each construct, the appropriate plasmid was transfected into DHFR CHO deficient cells for eukaryotic expression as described in Kaufman R.J. (1990) Methods Enzymol. 185, 537-566.

Expression of CDH3 (human, mouse and chimeric constructs) on CHO cells was tested in a FACS assay using a mouse anti-human CDH3 IgG1 monoclonal antibody that is cross-reactive. Bound monoclonal antibody was detected using Fcy-PE pro

- 49 042775 mouse IgG. As a negative control, cells were incubated with PBS/2% FCS (fetal calf serum) instead of the first antibody. Samples were analyzed by flow cytometry. The results are shown in FIG. 3. Expression of human and macaque CDH3 on CHO cells (see Example 5) was detected using PE-conjugated R&D 861-P.

Example 2 Epitope clustering of mouse scFv fragments

Cells transfected with human or mouse CDH3 or human/mouse CDH3 chimeric molecules (see Example 1) were stained with a crude, undiluted periplasmic extract containing scFv binding to human/macaque CDH3.

Bound scFv molecules were detected with mouse anti-FLAG-M2 monoclonal antibody (1 μU/ml; 50 μl in PBS/2%FCS; Sigma F1804) followed by anti-mouse IgG Fcy-PE (1:100, 50 μl; Jackson Immunoresearch No. 115-116-071). All antibodies were diluted in PBS with 2% FCS. As a negative control, cells were incubated with PBS/2% FCS (fetal calf serum) instead of periplasmic extract. Samples were analyzed by flow cytometry. The results are shown in FIG. 4.

In particular, in FIG. 4A shows the binding domains that recognize the extracellular D1 domain of human CDH3, and more specifically the DB subdomain (FACS signal loss in the respective CDH3 chimeric constructs). It should be noted that the binding domain, designated CDH3-6, is the parent binding domain for CDH3-4. The binding domain, designated CDH3-10, is the parental binding domain for CDH3-1, CDH3-2 and CDH3-3. In FIG. 4B shows the binding domains that recognize the extracellular D2 domain of human CDH3, and more specifically the D2C subdomain. It should be noted that the binding domain, designated CDH3-21, is the parental binding domain for CDH3-11, CDH3-12 and CDH3-14. The binding domain, designated CDH3-23, is the parental binding domain for CDH3-13. Finally, in FIG. 4C shows the binding domains that recognize the extracellular D3 domain of human CDH3, and more specifically the D3A subdomain. It should be noted that the CDH3-32 binding domain also binds to the D3C subdomain. The binding domain, designated CDH3-32, is the parental binding domain for CDH3-25, CDH3-26 and CDH3-27. The binding domain, designated CDH3-33, is the parental binding domain for CDH3-24. The term parent binding domain means in this context that these binding domains have been further developed in order to create or obtain optimized binding domains.

The binding domains CDH3-11, CDH3-12, CDH3-13, and CDH3-14 were also subjected to epitope clustering analysis and were shown to recognize the extracellular D2 domain of human CDH3, and more specifically the D2C subdomain (data not shown). The same analysis was also performed using the binding domains of CDH3-24, CDH3-25, CDH3-26 and CDH3-27. These binding domains recognized the extracellular D3 domain of human CDH3, and more specifically the D3A subdomain (data not shown).

Example 3 Biacore-Based Determination of Human and Macaque CDH3 Antibody Affinity

Biacore assay experiments were performed using recombinant CDH3 fusion proteins (human and macaque CDH3, respectively) with human albumin (HALB) to determine the binding of CDH3 targets to the antibodies of the invention.

Details CM5 Sensor Chips (GE Healthcare) were immobilized with approximately 600-800 RU of the appropriate recombinant antigen using pH 4.5 acetate buffer according to the manufacturer's guidelines. CDH3xCD3 bispecific antibody samples were loaded at five concentrations: 50 nM, 25 nM, 12.5 nM, 6.25 nM and 3.13 nM diluted in HBS-EP buffer (GE Healthcare). The flow rate was 30 μl/min for 3 min, then running buffer HBS-EP was applied again for 8-20 min at a flow rate of 30 μl/ml. Chip regeneration was performed using 10 mM glycine solution 10 mM NaCl pH 1.5. The datasets were analyzed using the BiaEval software. In general, two independent experiments were carried out. The CDH3xCD3 bispecific antibodies according to the invention showed high affinity for human CDH3 in the 1-digit nanomolar range. Binding to macaque CDH3 was balanced, also showing affinities in similar ranges. The affinity values, as well as the calculated affinity gap, are shown in Table. 2. CDH3-25 x F12q-HALB and CDH3-13- xI2C-HLE (Fc) were measured in a separate assay and were shown to have

KD (hu) of 31.8 ± 1.9 nM, KD (cyno) of 40.9 ± 3.89 nM and an affinity gap of 1.29, and

KD (hu) equal to 12.95 ± 0.5 nM, KD (cyno) equal to 12.35 ± 0.35 nM and affinity gap equal to 0.95, respectively.

- 50 042775

Table 2. Affinities of CDH3xCD3 bispecific antibodies to human and macaque CDH3 determined by Biacore analysis and calculated interspecies affinity gaps

Bispecific antibody CDH3xCD3 KD hu CDH3 [nM] KD cyno CDH3 [nM] Gap of affinity and cyno/hu CDH3-11 7.5 6.2 0.83 CDH3-12 8.8 7.5 0.85 CDH3-13 7.4 6.0 0.81 CDH3-14 9.5 8.1 0.85 CDH3-24 6.1 5.2 0.85 CDH3-25 7.9 6.7 0.85 CDH3-26 8.8 7.5 0.85 CDH3-27 8.7 7.6 0.87

Example 4 Scatchard-Based Specificity Analysis of the CDH3xCD3 Bispecific Anti-human and Macaque CDH3 Antibody on Target Antigen-Positive Cells and Determination of the Interspecies Affinity Gap

The affinities of CDH3xCD3 bispecific antibodies to CHO cells transfected with human or macaque CDH3 were also determined using the Scatchard assay as the most reliable method to measure the potential affinity gap between human and macaque CDH3. For Scatchard analysis, binding saturation experiments were performed using a monovalent detection system to accurately determine the monovalent binding of CDH3xCD3 bispecific antibodies to the respective cell line.

x 104 cells of the respective cell line (CHO cell line expressing recombinant human CDH3, CHO cell line expressing recombinant macaque CDH3) were each incubated with 50 µl of a triplet dilution series (twelve dilutions at 1:2) of the respective CDH3xCD3 (until saturation was reached), starting with 10-20 nM, followed by 16 h incubation at 4° C. with stirring and one step of washing the residue. The cells were then incubated for another hour with 30 μl of the CD3xALEXA488 conjugate solution. After one washing step, cells were resuspended in 150 µl of FACS buffer containing 3.5% formaldehyde, incubated for another 15 min, centrifuged, resuspended in FACS buffer and analyzed using a FACS CantoII machine and FACS Diva software. Data were obtained from two independent sets of experiments, each of which used three repetitions. The corresponding Scatchard analysis was calculated to extrapolate the maximum binding (Bmax). Concentrations of CDH3xCD3 bispecific antibodies were determined at half-maximal binding, reflecting the respective KDs. Triple measurements were presented as hyperbolic curves and as S-curves to demonstrate the correct concentration from minimal to optimal binding.

The values given in table. 3 were derived from two independent experiments on the CDH3xCD3 bispecific antibody. Cell-based Scatchard analysis confirmed that the CDH3xCD3 bispecific antibodies of the invention are nanomolar-subnanomolar in affinity for human CDH3 and present a small affinity gap, about 1, in the macaque/human cross-species CDH3. CDH325 x F12q-HALB and CDH3-13- xI2C-HLE (Fc) were measured in a separate assay and were shown to have a cell-based affinity (hu) of 0.37 ± 0.11 nM, cell-based affinity (cyno) of 0.35 ± 0.08 nM and an affinity gap of 0.95 and a cell-based affinity (hu) of 0.32 ± 0.003 nM, a cell-based affinity (cyno) of 0, 5 ± 0.09 nM and an affinity gap of 1.56, respectively

- 51 042775

Table 3. Affinities (KD) of CDH3xCD3 bispecific antibodies determined by Scatchard cell assay with calculated macaque KD CDH3/human CDH3 KD affinity gap.

Antibodies were measured in two independent experiments, each of which was used three times.

Bispecific CDH3xCD3 antibody Affinity for cell-based hu CDH-3 [nM] Affinity for cell basis macCDH-3 [nM] KDmac/KDhu CDH3 affinity gap CDH3-11 1.74±0.37 1.98±0.67 1.14 CDH3-12 0.40±0.27 0.52±0.21 1.30 CDH3-13 0.42±0.35 0.41 ± 0.28 0.98 CDH3-14 0.49±0.14 0.41 ± 0.03 0.84 CDH3-24 0.09±0.03 0.15±0.07 1.67 CDH3-25 0.12±0.01 0.11±0.03 0.96 CDH3-26 0.17±0.01 0.16±0.04 0.94 CDH3-27 0.20±0.10 0.23 ± 0.03 1.15

Example 5 Bispecific Binding and Cross-Species Cross-Reactivity

To confirm binding to human CDH3 and CD3 and macaque CDH3 and CD3, bispecific antibodies were tested by flow cytometry using CHO cells transfected with human and macaque CDH3 respectively, human CDH3 positive epidermoid carcinoma cell line A431, T cell line human leukemia HPB-all expressing CD3 (DSMZ, Braunschweig, ACC483); and the HSC-F T cell line expressing macaque CD3.

In addition, CHO cells transfected with mouse CDH3 were used as a negative control.

For flow cytometry, 200,000 cells of the respective cell lines were incubated for 30 min on ice with 50 μl of purified bispecific antibody at a concentration of 5 μg/ml. Cells were washed twice in PBS/2% FCS and binding of the constructs was detected with a mouse antibody (in-house) specific for the CD3 binding portion. After washing, bound mouse antibodies were detected using goat anti-mouse Fcy-PE. Samples were analyzed by flow cytometry.

The results are shown in FIG. 5A and 5B. The CDH3xCD3 bispecific antibodies of the invention stained CHO cells transfected with human CDH3 and macaque CDH3, and also stained the A431 epidermoid carcinoma cell line positive for human CDH3 (naturally expressing), as well as human and macaque T cells expressing CD3. In addition, there was no staining of negative control cells (CHO, CDH3 transfected mice).

Example 6 Confirmation of No Binding to Human and Macaque Paralogs

Human and macaque CDH3 paralogs (CDH1=E-cadherin, CDH2=N-cadherin, CDH4=R-cadherin and CDH5=VE-cadherin) were stably transfected into dhfr/CHO cells. Protein expression was confirmed in FACS assays with antibodies specific for the respective paralogs. Antibodies were R&D MAB18381 (for CDH1), eBioscience 12-3259-41 (for CDH2), R&D Systems polyclonal AF2217 (for CDH4) and BD Bioscience #555661 (for CDH5).

The paralog sequences used in this example are indicated in the sequence listing (SEQ ID NO: 41-44). They can also be found in the following GenBank attachment numbers:

NM_004360 [human CDH1]

NM_001792 [human CDH2]

NM_001794 [CDH4 human]

NM_001795 [human CDH5]

Flow cytometry analysis was performed as described in Example 5. The results are shown in FIG. 5C and 5D. The assay confirmed that none of the CDH3xCD3 bispecific antibodies of the invention cross-reacted with any of the human CDH3 paralogs tested.

The antibodies of the invention were also tested for cross-reactivity with the macaque CDH3 paralogs CDH1, CDH2, CDH4 and CDH5 (data not shown). Expression of the macaque paralog on CHO cells was tested with the same antibodies as described above for human paralogs. The monkey paralog sequences used in this example are listed in the sequence listing (SEQ ID NOS: 45-48). They can also be found in the following GenBank attachment numbers:

XM_002802516 [CDH1 macaque]

JU321883 [CDH2 macaque]

ХМ_002802511 [CDH5 macaque]

In addition, the macaque CDH4 sequence was obtained from the Ensembl Genome Browser (ENSM

- 52 042775

MUT00000017252) and fused at the N-terminus in frame with the human signal peptide CDH4 (amino acids 119).

Example 7 Human Germline Identity

To analyze the identity/sequence similarity of antibodies to human germline genes, the CDH3 binding domains of the invention were aligned as follows: Full VL, including all CDRs, were aligned; the full VH, including CDRs 1 and 2, but other than CDR3 was aligned with the human antibody germline (Vbase) genes. More details can be found in the description of this application. The results are presented below in table. 4.

Table 4. Identity of VH and VL to human germline

Bispecific antibody CDH3xCD3 CDH3-11 CDH3-12 CDH3-13 CDH3-14 Identity of VH and VL to the germline human [%] 92.7 93.6 87.8 94.1 CDH3-24 89.4 CDH3-25 89.8 CDH3-26 89.8 CDH3-27 90.7

Example 8 Cytotoxic Activity

The activity of the CDH3xCD3 bispecific antibodies of the invention in redirecting effector T cells against CDH3 expressing target cells was analyzed in five in vitro cytotoxicity assays.

The activity of CDH3xCD3 bispecific antibodies in redirecting stimulated human CD8+ effector T cells against human CDH3 transfected CHO cells was measured in an 18 hour ⁵¹ -chromium release assay. The results are shown in FIG. 6.

The activity of CDH3xCD3 bispecific antibodies in redirecting stimulated human CD8+ effector T cells against the CDH3-positive human carcinoma cell line A431 was measured in an 18 hour 51-chromium release assay. The results are shown in FIG. 7.

The activity of CDH3xCD3 bispecific antibodies in redirecting T cells in unstimulated human PBMCs against human CDH3-transfected CHO cells was measured in a 48-hour FACS-based cytotoxicity assay. The results are shown in FIG. 8.

The activity of CDH3xCD3 bispecific antibodies in redirecting T cells in unstimulated human PBMCs against CDH3-positive human carcinoma cell line A431 was measured in a 48-hour FACS-based cytotoxicity assay. The results are shown in FIG. 9.

To confirm that cross-reactive CDH3xCD3 bispecific antibodies are able to redirect macaque T cells against macaque CDH3 transfected CHO cells, a 48 hour FACS-based cytotoxicity assay was performed using a macaque T cell line as effector T cells. The results are shown in FIG. 10.

Example 8.1. Chromium release assay with stimulated human T cells

Stimulated T cells enriched for CD8 T cells were generated as described below. A Petri dish (diameter 145 mm, Greiner bio-one GmbH, Kremsmuenster) was coated with a commercially available specific anti-CD3 antibody (OKT3, Orthoclone) at a final concentration of 1 μg/ml for 1 h at 37°C. Unbound protein was removed with a single PBS wash. 3-5 x 10 ⁷ human PBMC were added to a pre-coated petri dish in 120 ml RPMI 1640 stabilized glutamine/10% FCS/IL-2 20 U/ml (Proleukin®, Chiron) and stimulated for 2 days. On the third day, the cells were harvested and washed once with RPMI 1640. 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 culture medium as above. CD8 cytotoxic T lymphocytes (CTLs) were enriched by depletion of CD4+ T cells and CD56+ NK cells using Dynal-Beads according to the manufacturer's protocol.

CHO target cells transfected with macaque CDH3 or human CDH3 were washed twice with PBS and 11.1 MBq ⁵¹ Cr labeled in a final volume of 100 μl RPMI with 50% FCS for 60 min at 37°C. The labeled target cells were then washed 3 times with 5 ml RPMI and then used in the cytotoxicity assay. The assay was performed on a 96-well plate in a total volume of 200 μl supplemented with RPMI with an E:T ratio of 10:1. An initial concentration of 0.01-1 μg/ml of purified bispecific antibody and 3-fold dilutions were used. The incubation time for the assay was 18 hours. Cytotoxicity was defined as the relative values of released chromium in the supernatant relative to the difference between maximum lysis (addition of Triton-X) and spontaneous lysis (no effector

- 53 042775 cells). All measurements were carried out in four repetitions. Chromium activity in supernatants was measured on a Wizard 3 gamma counter (Perkin Elmer Life Sciences GmbH, Cologne, Germany). Results analysis was performed using Prism 5 for Windows (version 5.0, GraphPad Software Inc., San Diego, CA, USA). EC ₅₀ values calculated by the analysis program from sigmoidal dose response curves were used to compare cytotoxic activity.

Example 8.2. Redirecting activity of stimulated human effector T cells against human CDH3-transfected CHO cells

The cytotoxic activity of the CDH3xCD3 bispecific antibodies of the invention was analyzed in a 51-chromium ( ⁵¹ Cr) releasing cytotoxicity assay using human CDH3-transfected CHO cells as target cells and stimulated human CD8+ T cells as effector cells.

The experiment was carried out as described in example 8.1.

The results are shown in FIG. biv tab. 5. The CDH3xCD3 bispecific antibodies exhibited potent cytotoxic activity against human CDH3 transfected CHO cells up to the 1-digit picomolar range. The claimed antibodies that are specific for the epitope cluster corresponding to human CDH3 positions 291-363 are present with a favorable epitope-activity ratio supporting strong cytotoxic activity mediated by the bispecific antibody. CDH3-25 x F12q-HALB and CDH3-13- xI2C-HLE (Fc) were measured in a separate assay and were shown to have an EC50 of 3.6 pM and 8.9 pM, respectively.

Table 5. EC ₅ o [pg/mL] values of CDH3xCD3 bispecific antibodies analyzed in a 51-chromium releasing cytotoxicity assay ( ⁵¹ Cr) using human CDH3-transfected CHO cells as target cells and stimulated human CD8 T cells _______c as effector cells _______

Bispecific antibody CDH3xCD3 EC50 [pg/ml] CDH3-11 76 CDH3-12 370 CDH3-13 7.2 CDH3-14 138 CDH3-24 14 CDH3-25 4.9 CDH3-26 21 CDH3-27 61

Example 8.3. Redirection activity of stimulated human effector T cells against CDH3-positive human carcinoma cell line A431

The cytotoxic activity of the CDH3xCD3 bispecific antibodies was analyzed in a 51-chromium ( ⁵¹ Cr) releasing cytotoxicity assay using the CDH3-positive A431 human carcinoma cell line as the source of target cells and stimulated human CD8+ T cells as effector cells. The analysis was carried out as described in example 8.1.

According to the results of 51-chromium release studies with stimulated human CD8+ enriched T-lymphocytes as effector cells and CHO cells transfected with human CDH3 as target cells, the CDH3xCD3 bispecific antibodies of the invention also strongly influence cytotoxic activity against naturally expressing cells. -targets (Fig. 7 and Table 6). CDH3-25 x F12q-HALB and CDH3-13- xI2C-HLE (Fc) were measured in a separate assay and were shown to have an EC ₅ o of 1.2 pM and 16 pM, respectively.

Table 6. EC ₅ o [pg/mL] values of CDH3xCD3 bispecific antibodies assayed in an 18-hour 51-chromium-releasing cytotoxicity assay ( ⁵¹ Cr) using the CDH3-positive A431 human carcinoma cell line as target cell source, and stimulated CD8 enriched T- human cells as effector cells

Bispecific antibody CDH3xCD3 EC50 [pg/ml] CDH3-11 47 CDH3-12 362

-54042775

Example 8.4. FACS-based cytotoxicity assay with unstimulated human PBMC

Isolation of effector cells

Human peripheral blood mononuclear cells (PBMC) were obtained by ficoll density gradient centrifugation from enriched preparations of lymphocytes (buffered sheaths), a by-product of blood banks that collect blood for transfusion. The buffy coats were supplied by the local blood bank and the PBMCs were prepared on the same day of blood collection. After ficoll gradient centrifugation and extensive washes with PBS Dulbecco (Gibco), remaining erythrocytes were removed from PBMC by incubation with erythrocyte lysis buffer (155 mM NH ₄ Cl, 10 mM KHCO 3 , 100 μM EDTA). Platelets were removed through the supernatant by centrifugation with PBMC at 100 x g. The remaining lymphocytes mainly cover B and T lymphocytes, NK cells and monocytes. PBMCs were maintained in culture at 37°C/5% CO2 in RPMI medium (Gibco) with 10% FCS (Gibco).

Depletion of CD14+ and CD56+ cells

CD14 microbeads (Milteny Biotec, MACS, no. 130050-201) were used to deplete CD14+ cells; human CD56 microbeads (MACS, no. 130-050-401) were used to deplete NK cells. PBMC were counted and centrifuged for 10 min at room temperature with 300 x g. The supernatant was discarded and the cell pellet was resuspended in MACS isolation buffer [80 μl/10 ⁷ cells; PBS (Invitrogen, no. 20012-043), 0.5% (v/v) FBS (Gibco, no. 10270-106), 2 mM EDTA (Sigma-Aldrich, no. E6511)]. CD14 microbeads and CD56 microbeads (20 μl/10 ⁷ cells) were added and incubated for 15 min at 4-8°C. Cells were washed with MACS isolation buffer (1-2 ml/10 cells). After centrifugation (see above), the supernatant was discarded and the cells were resuspended in MACS isolation buffer (500 ml/10 ⁸ cells). CD14/CD56 negative cells were then isolated using LS Columns (columns for liquid chromatography) (Miltenyi Biotec, no. 130-042-401). Cells PBMC wt./about. CD14+/CD56+ were cultured in complete RPMI medium, ie. RPMI1640 (Biochrom AG, no. FG1215) supplemented with 10% PBS (Biochrom AG, no. S0115), 1x essential amino acids (Biochrom AG, no. K0293), 10 mM Hepes buffer (Biochrom AG, no. L1613), 1 mM sodium pyruvate (Biochrom AG , no. L0473) and 100 U/ml penicillin/streptomycin (Biochrom AG, no. A2213) at 37°C in an incubator, until needed.

Target cell labeling

For cell lysis assays, flow cytometry assays used fluorescent membrane dye DiOC ₁₈ (DiO) (Molecular Probes, no. V22886) to label CHO cells transfected with human CDH3 or macaque CDH3 as target cells and distinguished them from effector cells. Briefly, cells were harvested, washed once with PBS and adjusted to 10 cells/ml in PBS containing 2% (v/v) FBS and the DiO membrane dye (5 μl/10 ⁶ cells). After incubation for 3 min at 37°C, the cells were washed twice in complete RPMI medium and the cell number was adjusted to 1.25 x 105 cells/ml. Cell viability was determined using 0.5% (v/v) EosinG isotonic solution (Roth, no. 45380).

Flow cytometric analysis

This assay was designed to quantify the lysis of CHO cells transfected with human CDH3 and macaque CDH3 in the presence of serial dilutions of CDH3 bispecific antibodies. Equal volumes of DiO-labeled target cells and effector cells (ie, PBMC w/v CD14+ cells) were mixed to give an E:T cell ratio of 10:1. 160 μl of this suspension was transferred to each well of a 96-well plate. 40 µl serial dilutions of CDH3xCD3 bispecific antibodies and negative control bispecific (a CD3-based bispecific antibody recognizing a non-relevant target antigen) or complete RPMI medium were added as an additional negative control. The cytotoxic reaction mediated by the bispecific antibody proceeded for 48 h in a humidified incubator with 7% CO2. The cells were then transferred to a new 96-well plate, and the integrity of the target cell membrane was monitored by adding propidium iodide (PI) at a final concentration of 1 μg/ml. PI is a membrane-impermeable dye that is normally eliminated from viable cells while dead cells accept it and become identifiable by fluorescent emission.

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

- 55 042775 ^ dead target cells

Cytotoxicity [%]= x100

P target cells n = number of events

Using GraphPad Prism 5 software (Graph Pad software, San Diego), percent cytotoxicity was plotted against respective bispecific antibody concentrations. Dose-response curves were analyzed using four parametric logistic regression models to evaluate sigmoid dose-response curves with a fixed hill slope and calculate an EC ₅₀ value.

Example 8.5. Redirecting activity of unstimulated human PBMCs against human CDH3-transfected CHO cells

The cytotoxic activity of CDH3xCD3 bispecific antibodies was analyzed in a FACS-based cytotoxicity assay using human CDH3-transfected CHO cells as target cells and unstimulated human PBMC as effector cells. The analysis was carried out as described in example 8.4 above.

The results of FACS-based cytotoxicity assays with unstimulated human PBMC as effector cells and CHO cells transfected with human CDH3 as targets are shown in FIG. 8 and in table. 7. CDH3-25 x F12q-HALB and CDH3-13- xI2C-HLE (Fc) were measured in a separate assay and were shown to have EC50s of 4.6 and 5.6 pM, respectively.

Table 7. EC50 [pg/mL] values of CDH3xCD3 bispecific antibodies analyzed in a 48-hour FACS-based assay using unstimulated PBMC cells as effector cells and CHO cells transfected with human CDH3 in . as target cells

Bispecific antibody EC50 [pg/ml] CDH3xCD3 CDH3-11 462 CDH3-12 1021 CDH3-13 129 CDH3-14 1885 CDH3-24 19 CDH3-25 10 CDH3-26 47 CDH3-27 61

Example 8.6. Redirecting activity of unstimulated PBMC cells against CDH3-positive human carcinoma cell line A431

The cytotoxic activity of the CDH3xCD3 bispecific antibodies was further analyzed in a FACS-based cytotoxicity assay using the CDH3-positive A431 human carcinoma cell line as target cell source and unstimulated human PBMC cells as effector cells. The analysis was carried out as described in example 8.4 above. The results are shown in FIG. 9 and in table. 8. CDH3-25 x F12q-HALB and CDH3-13- xI2C-HLE (Fc) were measured in a separate assay and were shown to have EC ₅₀ of 2.3 and 32 pM, respectively.

Table 8. EC ₅₀ [pg/mL] values of CDH3xCD3 bispecific antibodies analyzed in a 48-hour FACS-based assay using unstimulated PBMC cells as effector cells and human A431 cell line as target cells

It was expected that EC50 values were generally higher in cytotoxicity assays with unstimulated PBMCs as effector cells compared to cytotoxicity assays using stimulated human CD8+ T cells.

- 56 042775

Example 8.7. Redirecting activity of macaque T cells against macaque CDH3-transfected CHO cells

Finally, the cytotoxic activity of CDH3xCD3 bispecific antibodies was analyzed in a FACS-based cytotoxicity assay using macaque CDH3 (cyno) transfected CHO cells as target cells and a macaque T cell line as effector cells. The macaque T cell line 4119LnPx (Knappe et al. Blood 95:3256-61 (2000)) was used as a source of effector cells. Target cell labeling of CHO cells transfected with macaque CDH3 and analysis of cytotoxic activity based on flow cytometry were performed as described above.

The results are shown in FIG. 10 and in table. 9. Macaque T cells from the 4119LnPx cell line were induced to effectively kill macaque CDH3-transfected CHO cells with the CDH3xCD3 bispecific antibodies of the invention, that is, antibodies that bind to the human CDH3 epitope cassette corresponding to human CDH3 positions 291-363, and cover adjacent subdomains D2C (positions 291-327) and D3A (positions 328-363). The antibodies presented in this assay with EC50 values ranging from 2 digits to very low 4 digits pg/mL, suggesting that these antibodies are highly active in the macaque system.

Another group of anti-CDH3 antibodies were identified during epitope clustering (see example 2) that bind to the extracellular domain of D1 and more specifically to the D1B subdomain of human CDH3. Surprisingly, CDH3xCD3 bispecific antibodies of this group, although highly influencing cytotoxic activity against human CDH3-transfected CHO cells, showed very little cytotoxic activity against macaque CDH3-transfected CHO cells (see Fig. 10C and Table 9). Antibodies in this group showed significantly weaker activity with EC50 values in the very high 4-digit and even 5-digit pg/ml range. CDH3-25 x F12q-HALB and CDH3-13-xI2C-HLE (Fc) were measured in a separate assay and were shown to have EC50s of 1.4 pM and 4.0 pM, respectively. Thus, CDH3xCD3 antibodies of the invention that bind to the CDH3 epitope cassette corresponding to positions 291-363 were about 5 to about 1000 times more potent in the macaque system than antibodies that bind to the extracellular domain of D1, and more specifically, to subdomain D1B CDH3.

Table 9. EC50 values [pg/mL] of CDH3xCD3 bispecific antibodies that bind to the CDH3 epitope cassette corresponding to positions 291-363 (lines 1-8) and CDH3xCD3 bispecific antibodies that bind to the CDH3 epitope cassette/D1B subdomain (lines 9- 12) measured in a 48-hour FACS-based cytotoxicity assay with macaque 4119LnPx cell line as effector cells and macaque CDH3-transfected CHO cells as target cells.

Bispecific antibody CDH3xCD3 EC50 [pg/ml] 1 CDH3-11 1183 2 CDH3-12 1325 3 CDH3-13 257 4 CDH3-14 1015 5 CDH3-24 32 6 CDH3-25 14 7 CDH3-26 37 8 CDH3-27 98 9 CDH3-1 13055 10 CDH3-2 12862 eleven CDH3-3 8390 12 CDH3-4 8795

Example 9 Monomer-Dimer Conversion After (i) Three Freeze/Thaw Cycles and (ii) 7 Days of Incubation at 250 µg/mL

The CDH3xCD3 bispecific antibody monomer was subjected to various stress conditions followed by high throughput SEC to determine the percentage of initially monomeric antibody that was converted to a dimeric antibody, (i) 15 μg monomeric antibody was adjusted to a concentration of 250 μg/ml with total formulation buffer and then frozen at -80°C for 30 min followed by defrosting for 30 min at room temperature. After three freeze/thaw cycles, the dimer content was determined using HP-SEC;

(ii) 15 μg of monomeric antibody was adjusted to a concentration of 250 μg/ml with total formulation buffer, followed by incubation at 37° C. for 7 days. The dimer content was determined at

- 57 042775 power of HP-SEC.

A high resolution SEC TSK Gel G3000 SWXL column (Tosoh, Tokyo, Japan) was connected to an Akta Purifier 10 FPLC (GE Lifesciences) equipped with an A905 Autosampler. Balance buffer and running buffer consisted of 100 mm KH ₂ RHO ₄ - 200 mm Na ₂ SO ₄ brought to pH 6.6. The antibody solution (15 μg of protein) is applied to a balanced column, and the elution is carried out at a flow rate of 0.75 ml/min at a maximum pressure of 7 MPa. The entire run was monitored at optical absorbances of 280, 254 and 210 nm. The analysis was performed by peak integration of the 210 nm signal recorded in the Akta Unicorn software evaluation sheet. The dimer content was calculated by dividing the dimer peak area by the total monomer area plus the dimer peak.

The results are presented below in table. 10. CDH3xCD3 bispecific antibodies binding to the epitope cassette/extracellular subdomain of D2C are presented with <1% dimer content, more specifically, 0.0% dimer fractions after three freeze/thaw cycles and also after 7 days of incubation at 37 °C, which is considered very good. The conversion coefficients of the CDH3xCD3 dimer of bispecific antibodies of the epitope cluster/extracellular subdomain D3 A reached values <2%, more precisely from 0.2 to 1.8, which is considered good. CDH3-25 x F12q-HALB and CDH3-13 x I2C-HLE (Fc) were measured in a separate assay and were shown to have a percent dimer after three freeze/thaw cycles of 1.1 and 0.84, respectively, and a percent dimer after 7 days of incubation 0.0 (both HLE constructs).

Table 10. Percentage of Monomeric and Dimeric CDH3xCD3 Bispecific Antibodies Detected by High Performance Size Exclusion Chromatography (HP-SEC)

CDH3xCD3 antibody Percent dimer after three freeze/thaw cycles Percentage of dimer after 7 days of incubation CDH3-11 0.00 0.00 CDH3-12 0.00 0.00 CDH3-13 0.00 0.00 CDH3-14 0.00 0.00 CDH3-24 1.01 0.20 CDH3-25 1.31 0.60 CDH3-26 0.82 1.80 CDH3-27 1.69 1.50

Example 10 Thermal Stability

The antibody aggregation temperature was determined as follows: 40 μl of the antibody solution at 250 μg/ml was transferred into a disposable cuvette and placed in a Wyatt DynaPro Nanostar dynamic light scattering device (Wyatt). The sample was heated from 40 to 70° C. at a heating rate of 0.5° C./min, continuously obtaining a measured radius. An increase in radius indicating protein melting and aggregation was used by the software package supplied with the DLS device to calculate the antibody aggregation temperature.

All CDH3xCD3 bispecific antibodies tested according to the invention showed very favorable thermal stability at aggregation temperatures above 54° C., as shown in Table 1. 11 below. CDH3-25 x F12q-HALB was measured in a separate assay and was shown to have a thermal stability of 56.3°C.

Table 11. Thermal stability of bispecific antibodies determined by DLS ___________ (dynamic light scattering) ___________

Bispecific antibody CDH3xCD3 Thermal stability (aggregation DLS °C) CDH3-11 59.8 CDH3-12 55.9 CDH3-13 59.6 CDH3-14 59.6 CDH3-24 55.1 CDH3-25 55.4 CDH3-26 54.9 CDH3-27 54.1

Example 11 Stability after incubation for 24 hours in human plasma

Purified bispecific antibodies were incubated at a ratio of 1:5 in pooled human plasma at 37°C for 24 h-96 h at a final concentration of 2-20 μg/ml. After incubation in plasma, anti

-58042775 bodies were compared in a 51-chromium release assay with stimulated human T cells and CDH3-transfected CHO cells at an initial concentration of 0.01-0.1 μg/mL and with an effector cell:target cell (E:T) ratio of 10 :1 (assay as described in Example 8.1, chromium release assay with stimulated human T cells). Unincubated, freshly thawed bispecific antibodies were included as controls. The results are presented in table. 12. All antibodies tested had favorable plasma stability (EC ₅ q plasma/EC ₅ o control) < 4, the D3A binding antibody group even had plasma stability < 3. CDH3-25 x F12q-HALB and CDH313-xI2C -HLE (Fc) were measured in a separate assay and were shown to have an EC ₅₀ with plasma of 4.4 pM, an EC ₅₀ without plasma of 3.6 pM, and a plasma to control ratio of 1.2, and an EC ₅₀ with plasma of 3.4 pM, an EC ₅₀ without plasma of 8.9 pM, and a plasma to control ratio of 0.4, respectively.

Table 12 Antibody _EC50 values with and without plasma incubation and calculated plasma/control values

Bispecific antibody CDH3xCD3 EC ₅ o [pg/ml] with plasma without plasma Plasma to control (EC50 plasma/ECso control) CDH3-11 47 76 0.6 CDH3-12 296 370 0.8 CDH3-13 25 7.2 3.5 CDH3-14 291 138 2.1 CDH3-24 19 14 1.4 CDH3-25 9.9 4.9 2.0 CDH3-26 24 21 1.1 CDH3-27 129 61 2.1

Example 12 Turbidity at an antibody concentration of 2500 μg/ml ml of a solution of purified monomeric antibody 250 μg/ml was concentrated using rotary concentrators to 2500 μg/ml. After 16 hours storage at 5° C., the turbidity of the antibody solution was determined by optical absorbance OD340 nm against formulation buffer. The results are presented below in table. 13. All antibodies tested have a very suitable turbidity < 0.05. CDH3-25 x F12q-HALB and CDH3-13-xI2C-HLE (Fc) were measured in a separate assay and were shown to have turbidity at <2500 μg/ml of 0.066 and 0.026, respectively.

Table 13. Antibody turbidity after concentration to 2.5 mg/mL overnight

Bispecific antibody CDH3xCD3 Turbidity at 2500 µg/ml CDH3-11 0.035 CDH3-12 0.025 CDH3-13 0.030 CDH3-14 0.025 CDH3-24 0.019 CDH3-25 0.026 CDH3-26 0.028 CDH3-27 0.022

Example 14 Therapeutic Efficacy of CDH3xCD3 Bispecific Antibody in a Human Tumor Xenograft Model

On day 1 of the study, human epidermoid carcinoma cell line A-431 and freshly isolated human PBMC were subcutaneously injected into the right flank of female NOD/SCID mice (E:T cell ratio 1:2). Placebo control group 1 mice (n=5) received no effector cells and were used as non-transplant controls for comparison with placebo control group 2 (n=10 receiving effector cells) to monitor the effect of PBMCs on tumor growth in the absence of antibodies.

Mice were treated with 0.5 mg/kg/day (group 3, η = 10), 0.05 mg/kg/day (group 4, η = 10) or 0.005 mg/kg/day (group 5, η = 10) a CDH3xCD3 bispecific antibody that specifically binds to the CDH3 D2C extracellular subdomain by daily intravenous bolus injection for 10 consecutive days, starting approximately 2 hours after tumor cell injection on day 1.

Tumors were measured with calipers during the study and the results were evaluated

Claims (25)

by intergroup comparison of tumor volumes (TV). Tumor growth inhibition T/C [%] was determined by calculating TV as T/C% = 100 x (median TV of the analyzed group) / (median TV of the control group 2). The results are shown in FIG. 11. Treatment of mice with the CDH3xCD3 bispecific antibody at a dose of 0.5 mg/kg/day resulted in complete inhibition of tumor formation, and none of the animals developed a tumor until the end of the study (day 40). Example 15 Antitumor Activity of the HLE CDH3xCD3 Bispecific Antibody in the HCT-116 Tumor Model The assay was performed in male NOD/SCID mice injected subcutaneously with HCT-116 human colon carcinoma cells. Effector cells were in vitro expanded and activated human CD3 T cells (day 12). Treatment was started when the tumors reached a volume of ~200 mm (day 17). The control group was a placebo-treated q5d T cell group. The antibody having SEQ ID NO: 425 was administered at a concentration of 5 mg/kg/administration (group 2) and 0.5 mg/kg/administration (group 3) every five days (q5d) via intravenous bolus injections. The results are shown in FIG. 12A and 12B. In particular, in FIG. 12B highlights the result as responders (7/10) and non-responders (3/10). Although the reason for the non-response in 30% of the animals is unclear, in FIG. 12B shows that administration of the extended half-life bispecific construct at a concentration of 5 mg/kg to responding animals results in a cessation of tumor growth from the time of antibody administration. Example 16 T Cell Activation Activation Assay Isolated PBMC from healthy human donors were cultured with increasing concentrations of CDH3-13~I2C or CDH3-13xI2C-HALB bispecific antibodies for 48 h (0.001 pM-20 μM serial dilutions). Expression of the CD69 activation marker on CD4+ and CD8+ T cells was determined by immunostaining and flow cytometry and antigen-specific mAb conjugates. The results are shown in FIG. 13 and discussed above. Example 17 Pharmacokinetic Study in Macaque Constructs with Extended Half-Life Female cynomolgus monkeys received an IV (intravenous) infusion over 60 minutes with 0.015 mg/kg extended half-life CDH3xCD3 antibody construct (volume 1 ml/kg) in buffer. The three HLE formats tested were P156, HALB and HALB variant 1 (see SEQ ID NOs: 437, 443 and 444), each fused to the C-terminus of the construct. The half-life of these HLE constructs was calculated based on and plasma concentration determined at four (P156) and six (HALB, HALB version 1) time points between 96 hours of subsequent administration and the end of the study. The estimated half-life of these HLE constructs in the macaque model was 57 h, 63-85 h and 68 h, respectively. These properties of PK confirm IV dosing in humans once a week. CLAIM 1. A bispecific single chain antibody construct comprising a first human binding domain that binds to a human CDH3 epitope cluster on the surface of a target cell, wherein the first binding domain also binds to macaque CDH3, and comprising a second binding domain that binds to human CD3 on the surface of a T cell , wherein the human CDH3 epitope cassette is at amino acid positions 291-363 (SEQ ID NO: 36) of human CDH3 as shown in SEQ ID NO: 1, wherein said binding domains have the following format: pairs of VH regions and VL regions are in single strand format antibodies (scFv). 2. An antibody construct according to claim 1, characterized in that the first binding domain binds to an epitope located at amino acid positions 291-327 (SEQ ID NO: 34) of human CDH3. 3. An antibody construct according to claim 1, characterized in that the first binding domain binds to an epitope located at amino acid positions 328-363 (SEQ ID NO: 35) of human CDH3. 4. An antibody construct according to claim 1, characterized in that the first binding domain also binds to an epitope located at amino acid positions 404-440 (SEQ ID NO: 390) of human CDH3. 5. The design of the antibody according to any one of claims 1 and 2, characterized in that the first binding domain contains a VH region containing CDR-H1, CDR-H2 and CDR-H3, and a VL region containing CDRL1, CDR-L2 and CDR- L3 selected from the group consisting of: a) CDR-H1 as shown in SEQ ID NO: 149, CDR-H2 as shown in SEQ ID NO: 150, CDR-H3 as shown in SEQ ID NO: 151, CDR-L1 as shown in SEQ ID NO: 152, CDR-L2 as shown in SEQ ID NO: 153 and CDR-L3 as shown in SEQ ID NO: 154; b) CDR-H1 as shown in SEQ ID NO: 159, CDR-H2 as shown in SEQ ID NO: 160, CDR-H3 as shown in SEQ ID NO: 161, CDR-L1 as shown in SEQ ID NO: 162, CDR-L2 as shown in SEQ ID NO: 163 and CDR-L3 as shown in SEQ ID NO: 164; - 60 042775 c) CDR-H1 as shown in SEQ ID NO: 169, CDR-H2 as shown in SEQ ID NO: 170, CDR-H3 as shown in SEQ ID NO: 171, CDR-L1 as shown in SEQ ID NO: 172, CDR-L2 as shown in SEQ ID NO: 173 and CDR-L3 as shown in SEQ ID NO: 174; d) CDR-H1 as shown in SEQ ID NO: 179, CDR-H2 as shown in SEQ ID NO: 180, CDR-H3 as shown in SEQ ID NO: 181, CDR-L1 as shown in SEQ ID NO: 182, CDR-L2 as shown in SEQ ID NO: 183 and CDR-L3 as shown in SEQ ID NO: 184; e) CDR-H1 as shown in SEQ ID NO: 189, CDR-H2 as shown in SEQ ID NO: 190, CDR-H3 as shown in SEQ ID NO: 191, CDR-L1 as shown in SEQ ID NO: 192, CDR-L2 as shown in SEQ ID NO: 193 and CDR-L3 as shown in SEQ ID NO: 194; f) CDR-H1 as shown in SEQ ID NO: 199, CDR-H2 as shown in SEQ ID NO: 200, CDR-H3 as shown in SEQ ID NO: 201, CDR-L1 as shown in SEQ ID NO: 202, CDR-L2 as shown in SEQ ID NO: 203 and CDR-L3 as shown in SEQ ID NO: 204; g) CDR-H1 as shown in SEQ ID NO: 209, CDR-H2 as shown in SEQ ID NO: 210, CDR-H3 as shown in SEQ ID NO: 211, CDR-L1 as shown in SEQ ID NO: 212, CDR-L2 as shown in SEQ ID NO: 213 and CDR-L3 as shown in SEQ ID NO: 214; h) CDR-H1 as shown in SEQ ID NO: 219, CDR-H2 as shown in SEQ ID NO: 220, CDR-H3 as shown in SEQ ID NO: 221, CDR-L1 as shown in SEQ ID NO: 222, CDR-L2 as shown in SEQ ID NO: 223 and CDR-L3 as shown in SEQ ID NO: 224; i) CDR-H1 as shown in SEQ ID NO: 229, CDR-H2 as shown in SEQ ID NO: 230, CDR-H3 as shown in SEQ ID NO: 231, CDR-L1 as shown in SEQ ID NO: 232, CDR-L2 as shown in SEQ ID NO: 233 and CDR-L3 as shown in SEQ ID NO: 234; And j) CDR-H1 as shown in SEQ ID NO: 239, CDR-H2 as shown in SEQ ID NO: 240, CDR-H3 as shown in SEQ ID NO: 241, CDR-L1 as shown in SEQ ID NO: 242, CDR-L2 as shown in SEQ ID NO: 243 and CDR-L3 as shown in SEQ ID NO: 244. 6. The design of the antibody according to any one of claims 1, 3 and 4, characterized in that the first binding domain contains a VH region containing CDR-H1, CDR-H2 and CDR-H3, and a VL region containing CDR-L1, CDR- L2 and CDR-L3 selected from the group consisting of: a) CDR-H1 as shown in SEQ ID NO: 279, CDR-H2 as shown in SEQ ID NO: 280, CDR-H3 as shown in SEQ ID NO: 281, CDR-L1 as shown in SEQ ID NO: 282, CDR-L2 as shown in SEQ ID NO: 283 and CDR-L3 as shown in SEQ ID NO: 284; b) CDR-H1 as shown in SEQ ID NO: 289, CDR-H2 as shown in SEQ ID NO: 290, CDR-H3 as shown in SEQ ID NO: 291, CDR-L1 as shown in SEQ ID NO: 292, CDR-L2 as shown in SEQ ID NO: 293 and CDR-L3 as shown in SEQ ID NO: 294; c) CDR-H1 as shown in SEQ ID NO: 299, CDR-H2 as shown in SEQ ID NO: 300, CDR-H3 as shown in SEQ ID NO: 301, CDR-L1 as shown in SEQ ID NO: 302, CDR-L2 as shown in SEQ ID NO: 303 and CDR-L3 as shown in SEQ ID NO: 304; d) CDR-H1 as shown in SEQ ID NO: 309, CDR-H2 as shown in SEQ ID NO: 310, CDR-H3 as shown in SEQ ID NO: 311, CDR-L1 as shown in SEQ ID NO: 312, CDR-L2 as shown in SEQ ID NO: 313 and CDR-L3 as shown in SEQ ID NO: 314; e) CDR-H1 as shown in SEQ ID NO: 319, CDR-H2 as shown in SEQ ID NO: 320, CDR-H3 as shown in SEQ ID NO: 321, CDR-L1 as shown in SEQ ID NO: 322, CDR-L2 as shown in SEQ ID NO: 323 and CDR-L3 as shown in SEQ ID NO: 324; f) CDR-H1 as shown in SEQ ID NO: 329, CDR-H2 as shown in SEQ ID NO: 330, CDR-H3 as shown in SEQ ID NO: 331, CDR-L1 as shown in SEQ ID NO: 332, CDR-L2 as shown in SEQ ID NO: 333 and CDR-L3 as shown in SEQ ID NO: 334; g) CDR-H1 as shown in SEQ ID NO: 339, CDR-H2 as shown in SEQ ID NO: 340, CDR-H3 as shown in SEQ ID NO: 341, CDR-L1 as shown in SEQ ID NO: 342, CDR-L2 as shown in SEQ ID NO: 343 and CDR-L3 as shown in SEQ ID NO: 344; And h) CDR-H1 as shown in SEQ ID NO: 349, CDR-H2 as shown in SEQ ID NO: 350, CDR-H3 as shown in SEQ ID NO: 351, CDR-L1 as shown in SEQ ID NO: 352, CDR-L2 as shown in SEQ ID NO: 353 and CDR-L3 as shown in SEQ ID NO: 354. 7. An antibody construct according to claim 5, wherein the first binding domain comprises a VH region selected from the group consisting of VH regions as shown in SEQ ID NO: 155, SEQ ID NO: 165, SEQ ID NO: 175, SEQ ID NO: 185, SEQ ID NO: 195, SEQ ID NO: 205, SEQ ID NO: 215, SEQ ID NO: 225, SEQ ID NO: 235, and SEQ ID NO: 245. 8. An antibody construct according to claim 5, wherein the first binding domain comprises a VL region selected from the group consisting of VL regions as shown in SEQ ID NO: 156, SEQ ID NO: 166, SEQ ID NO: 176, SEQ ID NO: 186, SEQ ID NO: 196, SEQ ID NO: 206, SEQ ID NO: 216, SEQ ID NO: 226, SEQ ID NO: 236, and SEQ ID NO: 246. 9. An antibody construct according to claim 6, wherein the first binding domain comprises a VH region selected from the group consisting of VH regions as shown in SEQ ID NO: 285, SEQ ID NO: 295, SEQ ID NO: 305, SEQ ID NO: 315, SEQ ID NO: 325, SEQ ID NO: 335, SEQ ID NO: 345 and SEQ ID NO: - 61 042775 355. 10. An antibody construct according to claim 6, wherein the first binding domain comprises a VL region selected from the group consisting of VL regions as shown in SEQ ID NO: 286, SEQ ID NO: 296, SEQ ID NO: 306, SEQ ID NO: 316, SEQ ID NO: 326, SEQ ID NO: 336, SEQ ID NO: 346 and SEQ ID NO: 356. 11. An antibody construct according to claim 5, 7 or 8, characterized in that the first binding domain comprises a VH region and a VL region selected from the group consisting of pairs of a VH region and a VL region as shown in SEQ ID NO: 155+156 , SEQ ID NO: 165+166, SEQ ID NO: 175+176, SEQ ID NO: 185+186, SEQ ID NO: 195+196, SEQ ID NO: 205+206, SEQ ID NO: 215+216, SEQ ID NO: 225+226, SEQ ID NO: 235+236 and SEQ ID NO: 245+246. 12. An antibody construct according to claim 6 or 9, wherein the first binding domain comprises a VH region and a VL region selected from the group consisting of pairs of a VH region and a VL region as shown in SEQ ID NO: 285+286, SEQ ID NO: 295+296, SEQ ID NO: 305+306, SEQ ID NO: 315+316, SEQ ID NO: 325+326, SEQ ID NO: 335+336, SEQ ID NO: 345+346 and SEQ ID NO : 355+356. 13. An antibody construct according to any of the preceding claims, characterized in that it is in a format selected from the group consisting of (scFv) ₂ -, single domain scFv-mAb (monoclonal antibody), diabody and oligomers of the above formats. 14. An antibody construct according to any one of claims 5, 7, 8, 11 or 13, characterized in that the first binding domain contains an amino acid sequence selected from the group consisting of those shown in SEQ ID NO: 157, SEQ ID NO : 167, SEQ ID NO: 177, SEQ ID NO: 187, SEQ ID NO: 197, SEQ ID NO: 207, SEQ ID NO: 217, SEQ ID NO: 227, SEQ ID NO: 237 and SEQ ID NO: 247 15. An antibody construct according to any one of claims 6, 9, 10, 12 or 13, characterized in that the first binding domain contains an amino acid sequence selected from the group consisting of those shown in SEQ ID NO: 287, SEQ ID NO : 297, SEQ ID NO: 307, SEQ ID NO: 317, SEQ ID NO: 327, SEQ ID NO: 337, SEQ ID NO: 347 and SEQ ID NO: 357. 16. An antibody construct according to any one of the preceding claims, wherein the second binding domain binds to human CD3 epsilon and Callithrix jacchus, Saguinus Oedipus, or Saimiri sciureus. 17. An antibody construct according to any one of claims 5, 7, 9, 11, 12, or 13, having an amino acid sequence selected from the group consisting of those shown in SEQ ID NO: 158, SEQ ID NO: 168, SEQ ID NO: 178, SEQ ID NO: 188, SEQ ID NO: 198, SEQ ID NO: 208, SEQ ID NO: 218, SEQ ID NO: 228, SEQ ID NO: 238 and SEQ ID NO: 248, or having an amino acid sequence selected from the group consisting of those shown in SEQ ID NO: 379, SEQ ID NO: 380, SEQ ID NO: 381, SEQ ID NO: 382, SEQ ID NO: 383, SEQ ID NO: 384, SEQ ID NO: 385, SEQ ID NO: 386, SEQ ID NO: 387, SEQ ID NO: 388, SEQ ID NO: 389, SEQ ID NO: 422, SEQ ID NO: 423, SEQ ID NO: 424, SEQ ID NO: 425, SEQ ID NO: 426 and SEQ ID NO: 427. 18. An antibody construct according to any one of claims 6, 8, 10, 11, 12, or 14, having an amino acid sequence selected from the group consisting of those shown in SEQ ID NO: 288, SEQ ID NO: 298, SEQ ID NO: 308, SEQ ID NO: 318, SEQ ID NO: 328, SEQ ID NO: 338, SEQ ID NO: 348 and SEQ ID NO: 358. 19. A polynucleotide encoding an antibody construct according to any one of the preceding claims. 20. A vector for transformation or transfection of a host cell containing a polynucleotide according to claim 19. 21. A host cell for expressing an antibody construct, transformed or transfected with the polynucleotide of claim 19 or the vector of claim 20. 22. A method for producing an antibody construct according to any one of claims 1 to 18, comprising culturing a host cell according to claim 21 under conditions that allow expression of the antibody construct according to any one of claims 1 to 18, and isolating the produced antibody construct from the culture. 23. A pharmaceutical composition containing an antibody construct according to any one of claims 1 to 18 or obtained by the method according to claim 22, and additionally containing one or more carriers, stabilizers, excipients, diluents, solubilizers, surfactants, emulsifiers, preservatives and / or adjuvants. 24. The use of an antibody construct according to any one of claims 1 to 18, or obtained by the method of claim 22, in the prevention or treatment of cancer. 25. Use of an antibody construct according to claim 24, wherein the cancer is selected from the group consisting of lung carcinoma, head and neck carcinoma, primary or secondary CNS tumor, primary or secondary brain tumor, primary CNS lymphoma, spinal axis tumors, brainstem glioma , pituitary adenomas, adrenocortical cancer, esophageal carcinoma, colon cancer, breast cancer, ovarian cancer, NSCLC (non-small cell lung cancer), SCLC (small cell lung cancer), endometrial cancer, cervical cancer, uterine cancer, transitional cell carcinoma, bone cancer, pancreatic cancer, skin cancer, skin or intraocular melanoma, liver cancer, bile duct cancer, gallbladder cancer, kidney cancer, rectal cancer, anal cancer, cancer of the same -