WO2003090513A2 - Bi-specific antigen-binding compositions and related methods - Google Patents

Abstract

This invention provides a composition of matter comprising a first antigen-binding moiety and a second antigen-binding moiety operably affixed to one another via a flexible linker moiety. This invention also provides related nucleic acids, host-vector systems, compositions and methods of polypeptide production. This invention further provides related methods of treating a subject afflicted with a disorder mediated by the presence of an abnormal cell, and kits for practicing same.

Description

BI-SPECIFIC ANTIGEN-BINDING COMPOSITIONS AND RELATED METHODS

This application claims priority of U.S. Serial No. 60/374,930, filed April 23, 2002, the contents of which are incorporated herein by reference.

Throughout this application, various references are cited. Disclosure of these references in their entirety is hereby incorporated by reference into this application to more fully describe the state of the art to which this invention pertains .

Background of the Invention

Defective immunity is responsible for tumor development in cancer patients. In order to use a patient's own immune system to fight cancer, a number of cell-based adoptive immunotherapy approaches have been tried (9, 11, 12, 20). These approaches include lymphokine-activated natural killer cells, tumor-infiltrating lymphocytes, auto-lymphocytes, activation of lymph node-draining T cells,' antigen-specific cytotoxic T lymphocytes, anti- CD3-activated T cells, anti-CD3/anti-CD28 co-activated T cells, and dendritic cells. Although these approaches have been informative, clinical responses have usually shown no effect because of the lack of specificity toward any particular tumor.

New strategies have therefore been developed to combine the specificity of antibodies with the cytotoxic capability of T cells. The bi-specific monoclonal antibody (BsAb) approach is one of the new adoptive immunotherapy strategies. A BsAb, in one embodiment, consists of two monoclonal antibodies (mAbs ) cross-linked through chemical heteroconjugation . The BsAb will therefore carry dual specific "arms"; one arm recognizing and specifically binding to a tumor-associated antigen (TAA) and the other one recognizing the CD3 receptor on T cells. When a BsAb bridges a T cell and a tumor cell, the armed T cell can bypass the major histocompatibility complex (MHC) restrictions and become a TAA-specific cytotoxic T lymphocyte (CTL) against tumor cells bearing the TAA. In vitro, these BsAbs have shown specific cytotoxicity against tumors (25) . In the treatment of cancer, BsAbs have improved human survival rates and eradicated tumors in animals (24) .

Her2/neu is a member of the epidermal growth factor receptor family of tyrosine kinases that is over- expressed in several cancers, including breast cancer (21) . A chemically heteroconjugated anti-CD3 x anti- HER2/neu BsAb was used to treat high-risk breast cancer (13, 21) and hormone refractory prostate cancer.

However, chemically heteroconjugated BsAbs have important clinical limitations (1, 22, 24, 26) . First, the urine- derived mAbs induce HAMA (human anti-mouse antibody) responses in nearly all patients (5) . Second, chemical heteroconjugation procedures are still inefficient and inconsistent. Third, the heterogeneous conjugation product contains a mixture of monomer, dimer and multimer. Finally, the large molecular weight (>300 kDa) of BsAbs may prevent rapid tumor penetration.

Advances in antibody engineering have made it possible to overcome these restrictions by constructing recombinant bi-specific antibodies (re-BsAbs) that contain only the single chain fragments of variable regions (scFv) of mAbs, but still produce the same effector responses against tumor cells as whole mAbs do (2, 22, 24-26) . The re-BsAbs offer several advantages over intact BsAbs . First, the smaller molecule size (30-50 kDa) allows higher penetration into solid tumor tissues. Second, the HAMA reactions are largely reduced due to the lack of an immunogenic Fc domain of Ab. Third, the process of producing highly purified protein is greatly simplified. Finally, the entire protein production procedure can be done on a commercial scale.

Despite the recent advances in bi-specific antibody technology, structural and functional limitations still remain.

Summary of the Invention

This invention provides a first composition of matter comprising a first antigen-binding moiety and a second antigen-binding moiety operably affixed to one another via a flexible linker moiety.

This invention also provides a polypeptide comprising the amino acid sequence set forth in Figures 20-1 to 20-15 (SEQ ID NO:2) .

This invention also provides a polypeptide comprising the amino acid sequence set forth in Figure 25 (SEQ ID NO: ) .

This invention further provides a nucleic acid encoding a polypeptide comprising a first antigen-binding moiety and a second antigen-binding moiety operably affixed to one another via a flexible linker moiety having a length of ■ at least 16 amino residues.

This invention further provides a host-vector system comprising a host cell transfected with the instant expression vector.

This invention further provides a method for producing a polypeptide comprising a first antigen-binding moiety and a second antigen-binding moiety operably affixed to one another via a linker moiety having a length of at least 16 amino residues, which method comprises (a) culturing the instant host-vector system under conditions permitting the expression of the polypeptide, and (b) recovering the polypeptide so expressed.

This invention further provides a second composition of matter comprising (a) the above-described composition and (b) a cell having on its surface the antigen to which the first antigen-binding moiety specifically binds .

This invention further provides a method for increasing the activity of a CD3+ cell comprising contacting the cell with the instant composition.

This invention further provides a method for treating a subject afflicted with a disorder mediated by the presence of an abnormal cell, comprising administering to the subject (a) an agent known to ameliorate the disorder via contact with the abnormal cell, and (b) the instant composition, wherein the first antigen-binding moiety specifically binds to an antigen present on the agent, and the second antigen-binding moiety specifically binds to an antigen present on the abnormal cell.

This invention further provides a method for treating a subject afflicted with a tumor comprising administering to the subject (a) Interleukin-2 (IL-2), (b) T cells, and

(c) the antibody designated E3Bi .

This invention further provides a kit for use in treating a subject afflicted with a disorder mediated by the presence of an abnormal cell, comprising (a) the instant composition, wherein the first antigen-binding moiety specifically binds to an antigen present on an agent known to ameliorate the disorder and the second antigen- binding moiety specifically binds to an antigen present on the abnormal cell, and (b) instructions for use.

This invention further provides a kit for use in treating a subject afflicted with -a disorder mediated by the presence of an abnormal cell, comprising (a) the first ^■ instant composition, and (b) the agent known to ameliorate the disorder.

Finally, this invention provides a kit for use in treating a subject afflicted with a tumor comprising (a) Interleukin-2 (IL-2), (b) T cells, (c) the antibody designated E3Bi, and (d) instructions for use.

Brief Description of the Figures

Figure 1

This Figure shows the over-expression of EpCA on tumor cell surfaces but not on normal epithelium. The EpCAM is over-expressed in MCF-7 breast cancer cells (middle) and colorectal cancer cells (left) , but not in HBS-100 normal breast epithelial cells (right) . Cells were stained with the GA733.2 Ab.

Figure 2

This Figure illustrates the relationship of a T cell carrying a ch-TCR with and without the hinge spacer (H) . When the scFv binds to a specific antigen on the tumor cell surface, the connected T cell signaling chain "Y" initiates the T cell activation that will produce non- MHC-restricted tumor-killing activity.

Figure 3 Day 14 ATCs were used for these experiments. T cell populations from both healthy donors and patients were either not transduced (T) or transduced with the empty retrovirus only (SAM) , with the retrovirus carrying the GA733.2-derived ch-TCR (GA) , or with GA plus a hinge (GAH) . The effector-to-target ratios are 5:1 for panels A and B, or 2.5 : 1 for panel 3C. ELISAs of IFN-γ and TNF-α were performed after 24 hr incubation. Supernatants (50 μl) were collected for ELISAs in triplicate. The target cell lysis was determined after incubation for 4 hr at 37°C by the ⁵¹Cr-release assay. Only data from healthy donors are shown in panel C because there is no different cytotoxicity observed using either patients' or normal donors' ATCs. Panels A and B show that cytokine production was increased by the hinge addition (GAH with a hinge and GA without) . Panel C shows that cytotoxicity was also increased by about two-fold in the GAH group.

Figure 4 The ch-TCR with a hinge (GAHγ-EN) shows greatly increased T cell aggregations with the tumor cells in comparison to the ch-TCR without the hinge (GAγ-EN) . These photographs were taken after co-cultivation of tumor cells and T cells for 4 hr at 37°C at an effector-to-target ratio of 2:1. The arrows point at tumor cells, LS174T. "T cell", non-transduced T cells plus tumor cells; "SAM-EN", T cells transduced with expression vector only without the gene of interest; "GAγ-EN", with the hinge; "EN", an internal ribosome entry site in the vector.

Figure 5

The cytotoxicity of ch-TCR-transduced T cells only occurs when they are exposed to EpCAM-positive tumor cells (LS174T) at an E:T ratio of 5:1 for 24 hr at 37°C. The SD is indicated in both panels. These data also demonstrate that there is no significant difference between using the γ- or ζ-chain as the ch-TCR signaling domain to induce the cytolytic function of these transduced T cells.

Figure 6

These photographs show that the BsAb-mediated aggregation of T cells and tumor cells is specific to the mAb used. Day 14 cultured ATCs armed with 50 ng of OKT3/9184 BsAb bind (aggregate) to MCF-7 (upper left) . There is no aggregation in three negative controls: ATCs armed with 50 ng of irrelevant BsAb (upper right), unarmed (lower left) , or armed with a mixture of 250 ng of non- conjugated OKT3 and 250 ng of non-conjugated 9184 (lower right). The effector-to-target ratio is 25:1. These photos were taken after 24 hr co-incubation of MCF-7 cells with the BsAb armed TC. The mAb 9184 is an anti- Her2/neu mAb.

Figure 7 This Figure demonstrates that as few as 5 ng BsAb per 1x10^s T cells can trigger the cytotoxicity mediated by armed T cells. This cytotoxicity assay was performed using MCF-7 cells . The data presented in this Figure are summarized from three experiments in three different donors. -This Figure shows composite titration curves for unarmed TCs and ATCs armed with 0.5, 5.0 and 50.0 ng of OKT3/9184 BsAb at effector-to-target ratios of between 5 and 25 to 1; unarmed (T) or armed with 0.5 (•) , 5.0 (■) , and 50 (♦) ng BsAb/lxlO⁶ ATCs/ml.

Figure 8

This Figure shows that 40% of mice treated with BsAb- armed ATCs survived. The BsAb, OKT3XT84.66, was used for this experiment. SCID mice received 3Gy of total body irradiation to eliminate NK cells to ensure engraftment of tumor cells. The mice received subcutaneous co- injections of armed or unarmed ATCs (20xl0^δ ATCs) along with CEA-positive LS174T tumor cells (lxlO⁶ cells) (Winn assay) . The control group only received tumor cells and no ATCs. All non-ATC mice died of tumor progression by day 15 with tumor size >22mm. On day 100, 40% of mice that received armed BsAb were still alive, while only 10% were alive in the group that only received un-ar ed ATCs .

Figure 9

This Figure shows the construction of E3Bi into pGlEN vector . VH-VLe, the scFv of GA 33 . 2 ; VH-VL3, the ScFv of OKT3 ; SD, splicing donor; SA, splicing acceptor ; His , 6xHis-tag; IRES , an internal ribosome entry site; neo^r, a neomycine phosphotransferase gene . Figure 10

This is an illustration of the re-BsAb, E3Bi . VL, variable light chain of mAb; VH, variable heavy chain of mAb; H, hinge.

Figure 11

This is an illustration of the recombinant bi-specific antibody, E3Bi, which binds the T cell receptor on a T cell and the tumor associated-antigen EpCAM on a tumor cell. Once this complex is formed, the T cell will be activated by the receptor-E3Bi binding, and will become cytotoxic and kill a tumor cell.

Figure 12

T cell aggregation is dependent on the E3Bi doses. E:T = 10:1, Day 15 ATCs, target = LS174T.

Figure 13 This Figure shows a cytotoxicity assay (⁵¹Cr release assay) of E3Bi-armed T cells. Target = LS174T, 16 hr assay.

Figure 14 This Figure shows IFN-γ production induced by different doses of E3Bi .

Figure 15

This Figure shows the cloning of a hinge to the 3' -end of EpCAM scFv.

Figure 16

This Figure shows the construction of OKT3 scFv. Figure 17

This Figure shows the assembly of E3 to pGlEN.

Figure 18

This Figure .shows the replacement of a hinge with GS- linker GGGGSGGGGSGGGGS .

Figure 19 This Figure shows a circular map of pGlEN-EH3.His .

Figures 20-1 to 20-15

The complete DNA sequences of E3Bi and its vector have been confirmed by DNA sequencing analysis . This DNA plasmid is called pGlEN-EH3.His . The completed DNA sequence of 8,078 base pairs (SEQ ID NO:l) and the corresponding amino acid sequence (SEQ ID NO: 2) are also shown. The scFv of GA733.2 starts at site 1,388, the hinge starts at site 2,169, and the scFv of OKT3 starts at site 2,358. The 6XHis tag starts at site 3,093.

Figure 21

This Figure shows the in vivo anti-tumor response of E3Bi in a tumor xenograft model by tumor growth delay. SCID- Beige mice bearing LS174T xenografts were treated intratumoral (IT) injections IL-2 (n=6) , or IL-2/ATC

(n=8), or IL-2/ATC/E3Bi (n=6) beginning when tumor volumes of mice reached approximately 0.5 cc. Tumor growth delay is reported as the mean number of days (±SD) for tumor volumes of mice from each treatment group to reach 2 cc.

p = 0.0034 is the probability by Kruskal-Wallis non- parametric analysis that tumor growth delay is the same for all treatment groups, p < 0.01 is the probability by Dunn's multiple comparison analysis that treatment with IL-2/ATC/E3Bi produces the same tumor growth delay in mice as treatment with IL-2 alone; p > 0.05 for IL-2/ATC alone .

Figure 22

This Figure shows the survival of LS174T cells from LS174T tumor xenografts excised from SC D-Beige mice 24 h after mice received treatment with: IL-2 (300 IU/injection i.t.) alone; IL-2 and ATC (7xl0⁷ cells/injection i.t.); or IL-2/ATC and low (1 mg/kg i.v.) or high dose (10 mg/kg i.v.) E3Bi. After excision, tumor cells were processed into single-cell suspensions and seeded into cultures in four concentrations with five replicates each. Cells were counted after 7 days. Results are represented as the mean (±SE) surviving fraction of cells from each treatment group compared to the IL-2 treatment group. p < 0.001, IL-2 or IL-2/ATC vs. IL-2/ATC/E3Bi (10 mg/kg); p < 0.001, IL-2/ATC vs. IL- 2/ATC/E3Bi (1 mg/kg); p < 0.05, IL-2/ATC/E3Bi (1 mg/kg) vs. IL-2/ATC/E3Bi (10 mg/kg).

Figure 23

This Figure shows that E3Bi significantly triggers the cytotoxicity of PBMC (p = 0.0088). 1, 2, and 3 day cytotoxicity assays (CML) were conducted on PBMC as the effectors and LS174T colon tumor cells as target cells. The E/T ratio is 5. 100 pmole E3Bi/10⁶ effectors were used. The error bars show the standard deviations from the triplicate. This Figure also shows that there was some non-MHC restricted and non-specific cytolytic activity of T cells in E3Bi- group; however, this cytolytic activity is insignificant, p > 0.05. Figure 24

The cDNA sequence of E3Bi (SEQ ID NO:3).

Figure 25 The protein sequence of E3Bi (SEQ ID N0:4).

Figures 26-1 to 26-5

Alternative protein sequence version of pGlEN-EH3.His (SEQ ID NO:5). The completed DNA sequence of 8,078 base pairs (SEQ ID NO: 1) is also shown.

Detailed Description of the Invention

Definitions

As used in this application, except as otherwise expressly provided herein, each of the following terms shall have the meaning set forth below.

"Activated T Cell," also referred to herein as "ATC," shall have the meaning normally ascribed to it in the art. Characteristics of ATC include, without limitation, resumption of cell cycle, elevated IL-2 secretion, upregulated IL-2 receptor expression, limited proliferation, and differentiation into effector cells.

"Administering" shall mean delivering in a manner which is effected or performed using any of the various methods and delivery systems known to those skilled in the art. Administering can be performed, for example, intravenously, pericardially, orally, via implant, trans- mucosally, transdermally, intramuscularly, subcutaneously, intraperitoneally, intrathecally, intra- lymphatically, intralesionally, or epidurally. Administering can be performed, for example, once, a plurality of times, and/or over one or more extended periods .

The term "antibody" includes, by way of example, both naturally occurring antibodies (e.g., IgG, IgM, IgE and IgA) and non-naturally occurring antibodies. The term

"antibody" also includes polyclonal and monoclonal antibodies, and fragments thereof (e.g., antigen-binding portions) . Furthermore, the term "antibody" includes chimeric antibodies, wholly synthetic antibodies, human antibodies, humanized antibodies, and fragments thereof. "BsAb", also referred to herein as "bi-specific antibody", shall include, without limitation, a composition of matter comprising two operably affixed moieties, wherein each moiety is capable of binding to an antigen and comprises an antibody. BsAbs include, for example, (i) compositions comprising whole antibodies tethered together, (ii) single antibodies having two antigen-binding domains, each specific for a different antigen, (iii) single chain polypeptides, each comprising two antigen-binding domains linked via a region of at least 16 amino acid residues, and (iv) compositions comprising antigen-binding portions of antibodies operably affixed via chemical linkers.

"E3Bi" in this application is equivalent to "E3-Bi" found in the priority application.

"Flexible linker moiety" shall mean any chemical or biochemical moiety which (i) joins two antigen-binding moieties, (ii) comprises at least one chemical bond about which rotation is permitted, and (iii) permits the unhindered binding of each antigen-binding moiety joined thereto to its respective antigen. In the preferred embodiment, the flexible linker moiety permits binding of the two antigen-binding moieties to their respective antigens located on different cells (e.g., permitting the first antigen-binding moiety to bind to its antigen on a tumor cell, and the second antigen-binding moiety to bind to its antigen on a T cell) .

"Host cells" include, but are not limited to, bacterial cells, yeast cells, fungal cells, insect cells, and mammalian cells. Mammalian cells can be transfected by methods well-known in the art such as calcium phosphate precipitation, electroporation and microinjection.

"Mammalian cell" shall mean any mammalian cell. Mammalian cells include, without limitation, cells which are normal, abnormal and transformed, and are exemplified ^■ by neurons, epithelial cells, muscle cells, blood cells, immune cells, stem cells, osteocytes, endothelial cells and blast cells .

"Non-activated T cell" shall have the meaning normally ascribed to it in the art. Characteristics of a non- activated T cell include, without limitation, quiescence of cell cycle, non-proliferation and non-differentiation.

The terms "nucleic acid", "polynucleotide" and "nucleic acid sequence" are used interchangeably herein, and each refers to a polymer of deoxyribonucleotides and/or ribonucleotides . The deoxyribonucleotides and ribonucleotides can be naturally occurring or synthetic analogues thereof.

"Pharmaceutically acceptable carriers" are well known to those skilled in the art and include, but are not limited to, 0.01-0.1 M and preferably 0.05 M phosphate buffer or 0.8% saline. Additionally, such pharmaceutically acceptable carriers can be aqueous or non-aqueous solutions, suspensions, and emulsions. Examples of non- aqueous solvents are propylene glycol, polyethylene glycol, vegetable oils such as olive oil, and injectable organic esters such as ethyl oleate . Aqueous carriers include water, alcoholic/aqueous solutions, emulsions and suspensions, including saline and buffered media. Parenteral vehicles include sodium chloride solution, Ringer's dextrose, dextrose and sodium chloride, lactated Ringer's and fixed oils. Intravenous vehicles include fluid and nutrient replenishers, electrolyte replenishers such as those based on Ringer's dextrose, and the like.

Preservatives and other additives may also be present, such as, for example, antimicrobials, antioxidants, chelating agents, inert gases, and the like.

The terms "polypeptide," "peptide" and "protein" are used interchangeably herein, and each means a polymer of amino acid residues . The amino acid residues can be naturally occurring or chemical analogues thereof. Polypeptides, peptides and proteins can also include modifications such as glycosylation, lipid attachment, sulfation, hydroxylation, and ADP-ribosylation.

"Specifically bind" shall mean that, with respect to the binding of an antigen-binding moiety to its respective antigen, the moiety binds to that antigen with a greater affinity than that with which it binds to most or all other antigens. In the preferred embodiment, the moiety binds to that antigen with a greater affinity than that with which it binds to all other antigens.

"Stem cell" shall mean, without limitation, a cell that gives rise to a lineage of progeny cells . Examples of stem cells include CD34+ cells and embryonic stem cells. Surface adhesion molecules present on stem cells include, without limitation, IL-3 receptor, IL-6 receptor, IL-11 receptor, c-kit, VLA-4, VLA-5, L-selectin, PECAM-1 and Beta-1 integrin.

"Subject" shall mean any animal, such as a mammal or a bird, including, without limitation, a cow, a horse, a sheep, a pig, a dog, a cat, a rodent such as a mouse or rat, a chicken, a turkey and a primate. In the preferred embodiment, the subject is a human being. "Vector" shall mean any nucleic acid vector known in the art. Such vectors include, but are not limited to, plasmid vectors, cos id vectors, and bacteriophage vectors .

Embodiments of the Invention

This invention provides a first composition of matter comprising a first antigen-binding moiety and a second antigen-binding moiety operably affixed to one another via a flexible linker moiety.

The flexible linker moiety can comprise, for example, a polymer or a polypeptide. In one embodiment, the polypeptide has a length of at least 16 amino acid residues. In another embodiment, the polypeptide has a length of between 16 amino acid residues and about 100 amino acid residues . In another embodiment, the polypeptide has a length of between 50 amino acid residues and about 75 amino acid residues. In a further embodiment, the polypeptide has a length of about 63 amino acid residues, and/or comprises all or a portion of an antibody hinge region (e.g., CD8α Ig hinge-like region) . Preferably, the polypeptide has the amino acid sequence encoded by nucleotides 2170-2358 shown in

Figures 20-1 to 20-15 (SEQ NO ID:1).

Preferably, in the first composition, the first and second antigen-binding moieties specifically bind to different antigens. In one embodiment, the first antigen-binding moiety specifically binds to a tumor cell surface antigen. In another embodiment, the first antigen-binding moiety specifically binds to a cell surface antigen such as CD2, CD3, CD56 or other T cell or NK cell surface antigen. In a further embodiment, the first antigen-binding moiety specifically binds to a tumor cell surface antigen, and the second antigen- binding moiety specifically binds to a CD3+ cell surface antigen. In the preferred embodiment, the tumor cell surface antigen is EpCAM, and the CD3+ cell surface antigen is CD3. Other antigens include, for example, the breast cancer-associated antigen HER2. Antibodies against this antigen are known.

In another embodiment, the first antigen-binding moiety comprises the antigen-binding portion of an anti-EpCAM antibody, and the second antigen-binding moiety comprises the antigen-binding portion of the antibody designated OKT3. In another embodiment, the anti-EpCAM antibody comprises the antigen-binding portion of the antibody designated GA733.2.

In the first composition, each antigen-binding moiety preferably comprises the antigen-binding portion of an antibody. The antigen-binding portions can be, for example, Fab portions.

In one embodiment of the first composition, the composition comprises a single polypeptide chain which forms the first and second antigen-binding moieties and the linker moiety. In another embodiment, each of the first and second antigen-binding moieties further comprises a second polypeptide chain.

This invention further provides a polypeptide comprising the amino acid sequence set forth in Figures 20-1 to 20 15 (SEQ ID NO: 2) . This polypeptide is referred to herein as E3Bi, and comprises an anti-EpCAM and anti-CD3 domain. This invention further provides a polypeptide comprising the amino acid sequence set forth in Figure 25 (SEQ ID NO: 4 ) .

This invention further provides a nucleic acid encoding a polypeptide comprising a first antigen-binding moiety and a second antigen-binding moiety operably affixed to one another via a flexible linker moiety having a length of at least 16 amino residues. In one embodiment, the nucleic acid has the nucleotide sequence shown in Figures 20-1 to 20-15 (SEQ ID NO:l) . In another embodiment, the nucleic acid has the nucleotide sequence shown in Figure 24 (SEQ ID NO:3) .

The nucleic acid can be, for example, DNA or RNA, and is preferably DNA. In another embodiment, the nucleic acid is an expression vector. Expression vectors include, for example, plasmids, cosmids, bacteriophages and eukaryotic viruses. In one embodiment, the eukaryotic virus is an adenovirus or a retrovirus.

This invention further provides a host-vector system comprising a host cell transfected with the instant expression vector.

This invention further provides a method for producing a polypeptide comprising a first antigen-binding moiety and a second antigen-binding moiety operably affixed to one another via a flexible linker moiety having a length of at least 16 amino residues, which method comprises (a) culturing the instant host-vector system under conditions permitting the expression of the polypeptide, and (b) recovering the polypeptide so expressed. This invention further provides a second composition of matter comprising (a) the instant composition and (b) a cell having on its surface the antigen to which the first antigen-binding moiety specifically binds. In one embodiment, the cell is a CD3+ cell and the first antigen-binding moiety specifically binds to CD3.

In another embodiment, the cell is a T cell, the first antigen-binding moiety comprises the antigen-binding portion of the antibody designated OKT3, and the second antigen-binding moiety comprises the antigen-binding portion of the antibody designated GA733.2. In one embodiment, the composition of (a) is present in a ratio of from about 5-500 ng per million cells of (b) .

This invention further provides a method for increasing the activity of a CD3+ cell comprising contacting the cell with the first composition.

This invention further provides a method for treating a subject afflicted with a disorder mediated by the presence of an abnormal cell, comprising administering to the subject (a) an agent known to ameliorate the disorder via contact with the abnormal cell, and (b) the above- described composition, wherein the first antigen-binding moiety specifically binds to an antigen present on the agent, and the second antigen-binding moiety specifically binds to an antigen present on the abnormal cell.

In one embodiment, the subject is selected from the group consisting of a cow, a horse, a sheep, a pig, a dog, a cat, a rabbit and a primate. In the preferred embodiment, the subject is a human.

The disorder treated by the instant method can be any disorder mediated by an abnormal cell. Such disorders include, without limitation, cancer and specifically tumors. Cancer includes, without limitation, solid tumors, metastatic tumor cells and nonsolid cancers of the blood, marrow, and lymphatic systems. Tumors include, for example, carcinomas (cancers derived from epithelial cells), sarcomas (derived from mesenchymal tissues), lymphomas (solid tumors of lymphoid tissues) , and leukemias (marrow or blood borne tumors of lymphocytes or other hematopoietic cells) .

In a particular embodiment of the instant method, the agent is a CD3+ cell, the first antigen-binding moiety specifically binds to CD3 (or any other T cell antigen) , and the second antigen-binding moiety specifically binds to EpCAM. In another embodiment, the composition comprises the polypeptide whose amino acid sequence is shown in Figures 20-1 to 20-15 (SEQ ID NO: 2) . In another embodiment, the composition comprises the polypeptide whose amino acid- sequence is shown in Figure 25 (SEQ ID NO: 4) .

This invention further provides a method for treating a subject^' afflicted with a tumor comprising administering to the subject (a) Interleukin-2 (IL-2), (b) T cells, and (c) the antibody designated E3Bi . The T cells can be, for example, activated T cells or non-activated T cells.

In one embodiment, the subject is selected from the group consisting of a cow, a horse, a sheep, a pig, a dog, a cat, a rabbit and a primate. In the preferred embodiment, the subject is a human.

This invention further provides a kit for use in treating a subject afflicted with a disorder mediated by the presence of an abnormal cell, comprising (a) the first instant composition, wherein the first antigen-binding moiety specifically binds to an antigen present on an agent known to ameliorate the disorder and the second antigen-binding moiety specifically binds to an antigen present on the abnormal cells, and (b) instructions for use .

This invention further provides a kit for use in treating a subject afflicted with a disorder mediated by the presence of an abnormal cell, comprising (a) the first instant composition, and (b) the agent known to ameliorate the disorder. In one embodiment of the instant kits, the composition of (a) comprises a polypeptide having the sequence shown in Figures 20-1 to 20-15 (SEQ ID NO: 2) . In another embodiment of the instant kits, the composition of (a) comprises a polypeptide having the sequence shown in Figure 25 (SEQ ID NO:4) .

Finally, this invention provides a kit for use in treating a subject afflicted with a tumor comprising (a) Interleukin-2 (IL-2), (b) T cells, (c) the antibody designated E3Bi, and (d) instructions for use. The T cells can be, for example, activated T cells or non- activated T cells .

This invention will be better understood from the Experimental Details that follow. However, one skilled in the art will readily appreciate that the specific methods and results discussed are merely illustrative of the invention as described more fully in the claims that follow thereafter. Experimental Details

Introduction

The i munotherapeutic approach of using armed T cells with chemically conjugated bi-specific monoclonal antibodies (BsAbs) has shown specific cytotoxity against tumor cells. This BsAb carries dual specific "arms", one arm recognizing and specifically binding to a tumor associated antigen (TAA) , the other one to the CD3 receptor of T cells . When a BsAb bridges a T cell and a tumor cell, the armed T cell can bypass the major histocompatibility complex (MHC) restriction and become a TAA-specific cytotoxic T lymphocyte (CTL) . In the treatment of cancers, BsAbs have shown improvement of survival in humans and complete tumor eradication in animals .

Unfortunately, the use of these BsAbs is limited for long-term treatments for the following reasons. (1) Patients develop immune reactions against the BsAb because the BsAb was originally generated in mice. (2) The BsAb production is inconsistent from batch to batch. Using antibody engineering technologies, a genetically engineered recombinant BsAb (E3Bi) was constructed which contains only the sites for tumor and T cell binding but not the immunogenic site of the antibodies which causes unwanted reactions in patients . Generating highly purified protein products is greatly simplified and the entire procedure can be used in commercial production.

The TAA that E3Bi targets is called EpCAM (epithelial cell adhesion molecule) . EpCAM is over-expressed in all adenocarcinomas . Since EpCAM is a membrane protein and there is no soluble form in the serum to block antigen binding sites, and since EpCAM over-expressed in nearly all types of tumors, EpCAM was chosen as an ideal target for the E3Bi approach.

A re-BsAb was constructed from the mAb GA733.2 and mAb 0KT3, and called E3Bi. GA733.2 recognizes EpCAM (8).

The tumor targets

EpCAM (epithelial cell adhesion molecule, also called EGP-2, EGP-40, 17-1A, KSA) is a TAA that is over- expressed in all adenocarcinomas (23) . Since EpCAM is a membrane protein and there is no soluble form of it in the serum to block antigen binding sites, EpCAM is an ideal target for the re-BsAb approach. Figure 1 shows the surface antibody staining of EpCAM in colorectal (left) and breast (middle) cancer, as well as in normal epithelium (right) .

EpCAM is a well-studied and characterized tumor antigen. Two antibodies, C017-1A and GA733.2, bind to EpCAM, but at different epitopes and with different affinities . C017-1A has been used in clinical trials to treat colorectal cancer following surgery (6). However, there were no detectable immune responses reported. To direct a T cell to specifically target a TAA, the T cell receptor (TCR) can be engineered so that it carries the binding sites of a mAb that recognizes a TAA. This technique is also called a "T-body" or chimeric TCR (ch-TCR) approach. Daly et al . showed that only GA733.2 ch-TCR, not C017-1A ch-TCR, bound to EpCAM-positive tumor targets (4), probably because GA733.2 has a greater affinity than does

C017-1A (5xl0⁸ M^"1 compared with 0.7xl0⁸

respectively

(8)) . Preliminary Work

Addi tion of a hinge spa cer can significantly increase the tumor-binding and killing function of a ch-TCR

Two ch-TCRs (Figure 2) have been constructed. One has a hinge (H) insert and the other does not. Both ch-TCRs contain the scFv of the mAb GA733.2 that binds to the EpCAM and a T cell signaling domain that triggers T cell activation. Both the FcR γ-chain (GAHγ) and TCR ζ-chain (GAHζ) were used as the T cell signaling domain. This ch- TCR was transduced into an activated T cell (ATC) via a retrovirus . These results show that T cells carrying this ch-TCR specifically and efficiently target and lyse tumor cells (18), and the hinge spacer can increase the specific tumor lytic function (Figures 3 and 4) .

The addition of a hinge between the scFv and γ-chain greatly increased cytotoxic activity (18) (Figure 3). These results support the belief that the hinge spacer between these two scFv motifs improves binding efficiency to the targets as well as cytotoxicity.

Matthias Mack's group has designed a re-BsAb against EpCAM that is generated from a M79 hybridism (anti-17-lA) and they have shown its specific cytotoxicity in vi tro (10, 14, 15) . A 5-amino acid linker (G₄Sι) bridges these two scFvs in their respective re-BsAb. To date, they have not reported any results from in vivo experiments or clinical trials. However, their work has contributed to an improved design of re-BsAb: (1) the efficacy of the dual-headed recombinant antibody which contains only the scFv domains; (2) the feasibility of using mammalian CHO cells to express the fully functional recombinant protein; (3) the unnecessary addition of a co-stimulation portion in a re-BsAb construct; and (4) the stability of re-BsAb at 4°C for at least 6 months.

The mAb, GA733. 2, specifically binds to EpCAM- positive tumor cells, and not to EpCAM-nega tive cells

The cell line NCI-H716 is originally generated from cecum tumor cells that are EpCAM-negative. Using H716 as negative control, it was demonstrated that GA733.2 only targets EpCAM-positive cells (Figure 5) .

BsAbs are effective at specifically targeting tumor cells and genera ting cytolytic activi ty, and pre- arming T cells with BsAb before infusion increases the efficiency of BsAb-media ted tumor killing

To evaluate the potential efficacy of the E3Bi approach in future cancer therapy, the chemically heteroconjugated BsAb (OKT3/9184) targeting Her2/neu-positive breast cancer MCF-7 cells (21) was tested. In these experiments, there was a significant difference observed between adding the BsAb directly to the T cell and tumor cell mixture and pre-arming by adding the BsAb to T cells first, and then adding the armed T cells to the tumor targets (21) (data not shown) . Figure 6 shows, in the upper left panel, that activated T cells (ATCs) which had been cultured for 14 days and armed with 50 ng of BsAb (per 10⁶ cells) bind to and then kill MCF-7 cells. The lower left panel shows the ύn-armed ATC control. The ATCs in the upper right panel have been armed with irrelevant mAbs and those shown in the lower right have been armed with non-conjugated mAbs.

In order to determine the optimal arming dose for OKT3/9184, dose titration studies were performed at effector-to-target ratios of from 5:1 to 25:1. Increasing the arming doses led to increasing the mean percentage specific cytotoxicity. Figure 7 shows the specific cytolytic activity at different doses using ATCs from three healthy donors .

Using SCID mice and Winn assays (co-injection of tumor and T cells), a BsAb (OKT3xT84.66) was tested that specifically targets CEA (carcinoembryonic antigen) - positive colon cancer. The CEA-positive colorectal tumor cell line, LS174T, was used for these studies. Figure 8 shows that OKT3xT84.66 BsAb can prevent tumor progression and death in 40% of the mice.

Further Experiments

It is maintained that (1) the re-BsAb E3Bi, derived from mAb GA733.2, binds to EpCAM on tumor cells better than the re-BsAb from mAb C017-1A; (2) a hinge addition between two scFv motifs enhances the binding efficiency; and (3) pre-arming ATCs with the re-BsAb before infusion improves efficiency and minimizes clinical toxicity.

Purpose and Methods of Study

Specific Aims

To test this position, the following experiments were designed as set forth below.

Experiment 1: Construct E3Bi from two single chain fragments of variable regions (scFvs) of mAbs GA733.2 and 0KT3, and insert a linker from the CD8α hinge-like region (H) between these two scFvs. As a control, the H linker is replaced with a traditional glycine-serine linker, (G₃Sι)3. A 6xHis-tag is also constructed into the C- terminus of this recombinant protein for affinity purification purposes.

Experiment 2 : Express E3Bi in the mammalian cell line CHO and affinity purify E3Bi.

Experiment 3 : Evaluate the specific cytolytic function of E3Bi in vi tro (using EpCAM-positive colon cancer cell line LS174T, and using EpCAM-negative cecum cancer cell line H716 as a negative control) and in vivo (using Beige-SCID mice) .

Significance of the Instant Technology

The biggest challenge for cancer treatment is to direct a patient's own immune system to fight cancer. In general, tumor growth is the result of a defective immune system in which the MHC (major histocompatibility complex) fails to present tumor antigens to the immune system and to generate enough specific cytotoxic T lymphocytes (CTL) . Therefore, the adoptive immunotherapy strategies hold promise for cancer therapy because the focus of these treatments is to redirect a patient's own immune system to bypass MHC-restricted recognition and directly target tumor cells .

There are three major approaches in recently developed adoptive immunotherapy protocols. (1) Genetic modification of T cells to carry a chimeric T cell receptor (ch-TCR) that can recognize a specific tumor cell. Upon binding to the tumor cell, the ch-TCR will trigger the T cell to become a CTL and kill specific tumor cells. Because of the involvement of retrovirus production and gene transduction into ATCs in vitro, this treatment could be very expensive. (2) Dendritic cell (DC) -mediated tumor vaccination. Tumor antigens are introduced into DCs so that they can present these -tumor antigens to T cells and generate specific CTL. This strategy has not yet shown clinical success . Because there is no product that can be manufactured and specially trained medical technicians and facilities are required to perform this procedure, this treatment could also be very expensive and inconsistent. (3) Use of a conjugated bi-specific antibody (BsAb) molecule as a bridge between a tumor cell and a T cell so that the tumor cell will directly trigger the T cell to become a tumor-specific CTL.

Although the ch-TCR and DC approaches are important regarding proof-of-principle, they are very difficult, inconsistent and expensive to use in treating patients . Among these three approaches, the BsAb approach holds the greatest promise for clinical applications. It is technically feasible and straightforward.

The engineered recombinant BsAb approach (re-BsAb) overcomes the limitations of chemically heteroconjugated BsAbs because only the binding sites of the antibodies are selectively engineered, and not the regions that may cause side effects such as HAMA reactions. The re-BsAb product can be pure and consistent from lot to lot, while the chemically conjugated BsAb is only about 15-30% pure and the product is very inconsistent. The other advantage of this re-BsAb is that large-scale production is possible .

Again, this invention provides a re-BsAb with improved tumor-killing efficiency. This is accomplished in several ways: (1) using an antibody that has higher binding affinity; (2) adding a longer spacer between two binding sites; (3) producing this protein in mammalian cells; and (4) arming a patient's T cells with this re- BsAb in vi tro before infusing the patient.

Relevance to Cancer

Relapse rates remain unacceptably high after conventional treatments currently in use for solid tumors, like adjuvant chemotherapy/radiotherapy or even stem cell transplantation. There is an urgent need for nontoxic and tumor-specific approaches following both conventional and high dose chemotherapy to eradicate residual tumor cells and improve overall and disease-free survival. The goal of the E3Bi approach is to redirect a patient' s own immune system to specifically eradicate residual tumor cells following conventional treatments.

Because arming T cells with E3Bi will turn every T cell into a tumor-antigen specific CTL, E3Bi offers a very effective cancer immunotherapy approach. This product will have much less toxicity because the patient's own T cells will be stimulated to eradicate tumor cells . Pre- arming T cells before infusion will further increase the efficiency and specificity of this re-BsAb. Multiple infusions of these armed T cells over a longer period of time are expected to eradicate residual tumor cells more effectively compared to other immunotherapy approaches. The specificity of E3Bi is unique because these armed T cells will locally deposit at a specific tumor site and kill tumor cells. Furthermore, they will also attack residual tumor cells that have already spread prior to surgery.

Because colorectal cancer has the highest incidence among all types of cancer in the U.S., patients with colorectal cancer are envisioned as an important treatment group. Since EpCAM, the cell surface tumor marker recognized by

E3Bi, is over-expressed in all adenocarcinomas (23) , a very important aspect of E3Bi is that it has use with respect to most solid tumors as well.

It is expected that this re-BsAb will not only eradicate residual tumor cells, but will be part of adjuvant therapy for a variety of EpCAMt tumors.

Features of E3Bi

The design of E3Bi is unique and offers several advantages over other re-BsAbs that have been published (25) .

(i) The vector pGlEN is used for production of re-BsAb for the first time. Based on previous experience, this vector is highly effective in penetrating mammalian cell membranes, integrating cDNA into the host genome and promoting gene expression.

(ii) Mammalian cells (CHO cells) are used as E3Bi producer cells because mammalian proteins produced in the traditional bacterial cell E . coli may not fold properly and therefore may not function correctly.

(iii) The hinge spacer (63 amino acids) used for E3Bi has never been used for re-BsAb construction. The longer linker between two scFvs in E3Bi will provide the space needed for the interaction of a tumor cell and a T cell (18) and, therefore, is expected to increase the binding and tumor killing efficiency of E3Bi .

(iv) mAb GA733.2 is used for constructing a re-BsAb for the first time. Both GA733.2 and C017-1A target EpCAM, but at two different epitopes (7). GA733.2 has a higher affinity for EpCAM antigen than does C017-1A and produces stronger cytotoxicity against EpCAM-positive tumor cells

(4) . Increasing the affinity for a tumor antigen enhances the cytotoxicity of a bi-specific antibody.

(v) T cells from a patient are activated, expanded, armed with E3Bi and frozen for later infusion into the same patient. This in vitro arming protocol is the first of its kind used for a re-BsAb. It is believed that this approach not only provides a large quantity of activated and armed tumor-killing T cells, but also reduces the possible toxicity and increases the efficiency of E3Bi .

Detailed Experimental Methods

(1 ) Construction of E3Bi cDNA into a high expression vector

Figures 9-11 illustrate the design of E3Bi. (Also shown are the cloning of a hinge to the 3 '-end of EpCAM scFv (Figure 15), the construction of OKT3 scFv (Figure 16), the assembly of E3 to pGlEN (Figure 17), the replacement of a hinge with GS-linker GGGGSGGGGSGGGGS (Figure 18), and a circular map of pGlEN-EH3.His (Figure 19)).

The E3Bi cDNA is generated by combining variable light (V_L) and heavy (V_H) chains of mAbs GA733.2 and 0KT3 that are amplified by PCR. The E3Bi cDNA is then inserted into an expression vector, pGlEN. PGlEN is generated from the Maloney murine leukemia virus (MMLV) and is replication incompetent due to the lack of three genes that are essential for virus formation, gag, env and pol . This insures against retroviral replication. Based on previous experience (18), this vector is highly efficient in producing stably transduced mammalian cells and promoting gene expression. The anti-CD3 scFv is generated' from OKT3 hybridoma cells (ATCC, Rockville, MD) .

The E3Bi gene expression is driven by long terminal repeats (LTR) . This vector contains a leader sequence from the k light chain to penetrate cell membranes, a neomycin phosphotransferase gene (neo^r) for drug selection, a splicing donor (SD) /splicing acceptor (SA) to enhance the efficiency of transcription, and an internal ribosome entry site (IRES) for driving neo^r gene transcription. A 6xHis-tag is added to the C-terminal end for affinity purification of this re-BsAb protein. These two scFvs are linked through a long hinge that is cloned from the CD8α hinge-like region. By adding distance from the scFv to the plasma membrane, the hinge spacer has shown increased tumor binding and killing activity in connection with the chimeric TCR approach (16, 18) . Figure 3 shows that in both healthy donors and patients, the ch-TCR with a hinge (GAH) significantly increased the specific tumor cytotoxicity and cytokine secretion (IFN-γ and TNF- ) by about two-fold (compared to a ch-TCR without a hinge) (18) . However, the hinge approach has never before been applied for re-BsAb construction. The hinge is expected to give the re-BsAb flexibility and rotational freedom leading to a better bridge between a tumor cell and a T cell.

(2) Expression of E3Bi in eukaryotic cells

Most re-BsAbs are expressed in a traditional prokaryotic expression system (24). However, the re-BsAb protein may not fold properly in prokaryotic cells (14) . Therefore, a eukaryotic cell line, the Chinese hamster ovary cell line (CHO, GIBCO Life-technologies, Rockville, MD) , is transfected. Specifically, CHO is transfected with the standard CaP0₄ precipitation method (17) and cultured in the presence of the selection drug, G418. The stably transfected CHO cells form colonies after 10-14 days. The colonies are selected for the highest quantity of the re- BsAb production and evaluated by ELISA for the presence of a 6xHis-tag (Ni-NTA HisSorb Plates, QIAGEN, Valencia,

CA) . The re-BsAb is secreted into the culture medium that

, is used directly for functional evaluation without further purification. The CHO clone with the highest yield of re-BsAb is grown as non-adherent cells in a serum-free medium specially constituted for CHO (CD-CHO,

GIBCO) . The medium containing E3Bi is collected every 24 hr or as otherwise determined.

(3) Functional assays of E3Bi

(3. 1) In vi tro studies

Specific cytolytic and cytokine production assays are performed in both EpCAM-positive (LS174T from ATCC) and negative cells (H716 from ATCC) using the same techniques as described before (18, 21). Figure 8 demonstrates that, using anti-EpCAM mAb (GA733.2) staining, LS174T colorectal cells show very strong surface EpCAM expression.

For these in vi tro studies, T cells from healthy donors are isolated from 40 cc peripheral blood, activated with anti-CD3 mAb at 10-50 ng/lxlO⁶ T cells/ml, and expanded for 14 days in the presence of 100 IU of IL-2 and 10% fetal calf serum in the medium, RPMI-1640 (BioWittaker, Walkersville, MD) . On day 14, the ATCs are armed with different doses of E3Bi and rocked for 1 hr at 4°C. The cells are washed twice with RPMI-1640 to eliminate excess unbound E3Bi. The armed and unarmed ATCs are added to the target tumor cells at effector-to-target ratios from 1:1 to 10:1. Cytotoxicity assays (⁵¹Cr release assay) and IFN- γ production assays (ELISA) are performed in triplicate. The dose, time and temperature in the arming procedure are evaluated. To test the specific targeting of E3Bi against EpCAM, a blocking assay is performed using the anti-Id antibody against the scFv of GA733.2. The cytotoxicity and ELISA assays are analyzed statistically with a standard statistical package, a paired t-test or Wilcoxon signed tank test using the SigmaStat. All in vitro assays are repeated with at least 5 unrelated subjects. The significant cytotoxic functions of E3Bi are analyzed with a paired fc-test or Wilcoxon signed tank test using SigmaStat.

(3. 2) In vivo s tudies

In vivo functional assays are performed in animal models. Four to eight week-old female beige SCID mice are used for these studies (Taconic Pharm, Germantown, NY) . These mice carry the SCID mutation that causes a deficiency of both T and B cells resulting in cytotoxic T cell and macrophage defects as well as selective impairment of NK cell function. The animals are maintained in accordance with NIH animal care guidelines.

(3. 2. 1) Winn assay

The mice are divided into two groups; one group with "Winn Assay", which means 1x10^s tumor cells are co- injected with armed ATCs (dose range from 1x10^s to 10x10^s) subcutaneously into the upper right thigh of each animal or with unarmed T cells as a control. Tumor development is documented weekly. The other group is injected only with IxlO⁶ tumor cells subcutaneously into the upper right thigh of each animal. Once the tumor is established

(>5mm, about 4 weeks), armed or unarmed T cells at different doses are injected twice a week directly into the center of the tumor mass. As a control, mice from both groups are divided into three sub-groups: the first group receives no T cells; the second group is injected with unarmed T cells, and the third group is injected with armed T cells with E3Bi. The tumor development is measured and documented every 2 days. The tumor cells used for these in vivo studies are LS174T (human colorectal adenocarcinoma cells). T cells are extracted from the peripheral blood of both healthy donors and patients. The animals are sacrificed by C0₂ gas overdose once the tumor size exceeds 1.5 cm. By week 8-10 after treatment, all animals are sacrificed. All data are analyzed using a paired t-test or Wilcox test on signed tank test using SigmaStat.

(3. 2. 2) Tumor xenograft model - Xenografted mice wi th EpCAM+ tumor cells

The in vivo anti-tumor response of E3Bi was also evaluated in a tumor xenograft model by tumor growth delay assay. In SCID-Beige mice bearing xenografted LS174T tumors, the average time for tumors to reach four times their pre-treatment volume (0.5 cc) varied significantly between the following three treatment groups (p = 0.0034): animals treated by intratumoral (i.t.) injections with IL-2 alone; with IL-2 plus ATC; and with IL-2 plus E3Bi/ATC. Administration of ATC with IL-2 resulted in a tumor growth delay of 7 days compared to IL-2 treatment alone (p > 0.05), while addition of E3Bi to the treatment regimen significantly increased tumor growth delay by 12 days compared to IL-2 alone, (p < 0.01) . As shown in Figure 21, these results show that E3Bi significantly prolongs the survival rate of tumor-bearing mice, and therefore, provide a therapeutic advantage for using E3Bi with ATC/IL-2 to increase tumor growth inhibitions .

The same xenografted mouse model was also used to evaluate the trafficking and high does tolerance of parenterally-administered E3Bi in vivo . Four-week old SCID-Beige mice were divided into four groups: i.t. injection of IL-2 only (IxlO⁴ IU/kg) ; i.t. injection of IL-2 and ATC (2xl0⁹ cells/kg); i.v. injection of a low (1 mg/kg) or high (10 mg/kg) dose E3Bi along with an i.t. injection of IL-2/ATC. Each mouse received two i.v. injections (day 1 and day 3) of E3Bi and two i.t. injections of IL-2/ATC, day 1 with 14-day old ATC from a healthy donor (N4) and day 3 with 17-day old. Tumor necrosis was observed within 48 h after the injection in mice receiving high dose E3Bi, but not in mice receiving low dose E3Bi, ATC/IL-2 or IL-2 only. High dose E3Bi was well-tolerated with no evidence of any side effects.

The tumor size more than doubled in the mice receiving only ATC/IL-2 while it remained largely unchanged in mice receiving low dose E3Bi after 7 days from the. first injection. In addition to the observed necrosis (E) of tumors in mice receiving high dose of E3Bi, the tumors in these mice demonstrate partial regression within 7 days of initial treatment.

Figure 22 further supports the targeting specificity of E3Bi to EpCAM+ over-expressing tumors in vivo. Mice with established LS174T tumors were treated with ATC or ATC followed by an IV injection of low or high does E3Bi, and excised 24 h later. The viability of treated cells was measured as the surviving fraction of tumor cells after in vivo treatment with IL-2, IL-2 /ATC and IL-2/ATC/E3Bi . Though ATC treatment alone produced no cytotoxic effect on tumor cells, administration of low dose (1 mg/kg) E3Bi in conjunction with ATC treatment produced a 40% decrease in tumor cell survival. Increasing the E3Bi dose to 10 mg/kg significantly decreased the tumor cell survival by 90% (p < 0.05) . Combined with the tumor growth inhibition studies, these results show that E3Bi. delivered systematically traffics, binds and produces cytotoxic effects to EpCAM+ over-expressing tumor cells in vivo .

(3. 2. 3) E3Bi triggered cytotoxicity of non- activa ted T cell activa tion

E3Bi also directly triggers non-activating T-cells to kill tumor cells. For example, E3Bi triggered CD4+ and CD8+ populations in peripheral blood mononuclear cells (PBMC) to become activated in the presence of LS174T tumor cells (data not shown) . Both T cell activation markers, CD25 and CD69, increased upon activation by E3Bi and resulted in increased cytolytic activity of T-cells, as shown in Figure 23.

Figure 23 illustrates that E3Bi triggers cytotoxicity in PMBC, which include non-activated T cells. 1, 2, and 3 day cytotoxicity assays (CML) were conducted using PBMC as the effectors and LS174T colon tumor cells as target cells.. On day 3, the cytotoxicity of PBMC rose to 70%, and therefore, shows that E3Bi significantly triggers the cytotoxicity of PBMC (p = 0.0088). This Figure also shows some non-MHC restricted and non-specific cytolytic activity of T cells in the E3Bi- group; however, this cytolytic activity is insignificant, p > 0.05. (3. 3) Anticipated obstacles

(3. 3. 1) Clearance of E3Bi by kidney before i t can a ttack tumors

One major concern for a small sized re-BsAb is that it can be cleared rapidly by the kidney, and therefore, the amount of its retention by the tumor is very limited (3,

22) . To overcome this problem, T cells are pre-armed in vitro with E3Bi before infusion so that the small E3Bi will remain attached to the CD3 receptor on the T cells while traveling in the body and, therefore, be protected from rapid kidney clearance. More importantly, pre-arming the T cells in vitro will dramatically improve the killing efficiency (data not shown) . Existing methodology enables one to pre-arm T cells in vi tro for future clinical trials. The pre-arming procedure includes (1) mixing day 14 ATCs with different doses of E3Bi in a tube and rocking for one hour at 4°C; (2) washing twice to remove unbound E3Bi; and (3) infusing the armed T cells at a concentration of lxlO⁷ cells/ml.

(3. 3. 2) No costimula tion

Without CD28 costimulation, T cell activation can result in activation-induced T cell apoptosis (AICD) and as a consequence, reduced tumor killing efficiency in vivo . These phenomena have not been observed using ch-TCR (18, 19) and BsAb (10) . However, to confirm that there is no AICD in re-BsAb-mediated tumor killing activities and the armed ATCs can be recycled in vivo, bystander-killing assays, apoptosis assays (Annexin V staining) and ³[H]- thymidine proliferation assays are performed. Following tumor exposure, the fate of armed T cells is studied with and without the E3Bi. (4) Affinity purifica tion of E3Bi

The high producer cells are grown in suspension in the serum-free medium, CHO-S-SFM II (GIBCO) , which is a constituted medium developed specifically for CHO cells growing in suspension. The supernatant containing the released E3Bi is collected every 24 hr and affinity purified by applying it to Ni-NTA spin columns. These columns can purify up to 150 mg of E3Bi in a one-step affinity purification of 6xHis-tag-containing recombinant protein (QIAGEN) . The columns are washed and eluted according to the manufacturer's instructions. The quality of the purified product is evaluated by denaturing gel electrophoresis and Western blot. For the short term, E3Bi is stored in phosphate-buffered saline at 4°C, lyophilized and stored at -20°C for the long term.

(5) Affini ty purifica tion of E3Bi

Collected supernatant containing the E3Bi is applied to a Ni-NTA agarose column (nickel-charged resin, QIAGEN) . The concentration of eluted E3Bi is tested with the BCA testing kit (BCA-200 Protein Assay Kit, Bio-Rad, Hercules, CA) . The final product is filtered through a 0.22 mm filter, aliquoted in 1 mg protein/ml PBS/vial and stored in the -20°C freezer. This affinity purification is conducted in a cold box (4°C) in the GMP lab.

(6) Activa tion and expansion of T cells in vitro

It is routine to activate T cells in gas-permeable plastic bags with anti-CD3 antibody, OKT3 (OrthoBiotech, Raritan, NJ) . Briefly, T cells from healthy donors or patients are transferred into bags at a concentration of lxlO⁶ CD3⁺ cells/ml RPMI culture medium (BioWittaker) supplemented with 2-5% human serum, 100-500 IU of IL-2/ml and 20 ng OKT3/ml. T cells are maintained at a concentration of lxlO⁶ cells/ml for 14 days.

(7) Arming ^'activated T cells with E3Bi

The procedure for arming T cells with E3Bi is adopted from established procedures for using chemically heteroconjugated BsAb. Briefly, day 14 ATCs are transferred into a tube, washed and re-suspended in an optimal volume of culture medium containing the optimized dose of E3Bi . After incubation, excess E3Bi is washed twice by centrifugations . The armed ATCs are either aliquoted and frozen or directly used for functional studies.

(8) E3Bi-Media ted T Cell Killing

As shown in Figure 12, T cell aggregation is dependent on the E3Bi doses. Specifically, three photos show the binding of T cells (small round dots) and tumor cells (growing in "island-like" groups) mediated by E3Bi . The CHO cell culture supernatant that contains E3Bi was added to the T cell and tumor cell mixture. Panel A contains no CHO supernatant and there is no binding or aggregation between T cells and tumor cells. In panel B (12.5%), there are a significant number of T cells attached to the tumor cells. In panel C (25%), all tumor cells are aggregated with T cells. These panels clearly show that E3Bi can direct T cells to kill tumor cells. The concentration of E3Bi in the supernatant was not determined. As a control, the same CHO supernatant that contains recombinant protein other than E3Bi did not produce the same aggregation effects (data not shown) . Figure 13 shows a ⁵¹Cr release assay of E3Bi-armed T cells. This cytotoxicity assay shows the percentage of targets (tumor cells) that are killed by the effectors (T cells) in the presence of E3Bi. "E/T" indicates the number of T cells per tumor cell. These data show that at 16 hours, about 70% of tumor cells are killed at E/T =

10, and 50% at E/T = 5. Supernatant collected from 50% confluent E3Bi-transduced CHO cell culture was used for this assay. The "mock" is a control, wherein only an "empty vector" (i.e., without an E3Bi insert) was transduced into CHO cells and the supernatant was used.

As shown in Figure 14, IFN-γ production is induced by different doses of E3Bi . CHO cell culture supernatant containing secreted E3Bi was added to T cell and tumor cell mixtures at different doses in microliters as indicated. The absolute concentration of E3Bi was not determined. The cytotoxic function of T cells is usually indicated by the amount of their IFN-γ production. These data clearly show that E3Bi induces significant IFN-γ production in a dose-dependent manner, while the control group does not stimulate IFN-γ production.

References

1. Weiner, L.M., Bi-specific antibodies in cancer therapy. Cancer J. Sci. Am. 2000, 6 Suppl 3:S265- S271.

2. Bodey, B., et al . , Genetically engineered monoclonal antibodies for direct anti-neoplastic treatment and cancer cell specific delivery of chemotherapeutic agents. Curr. Pharm. Des 2000, 6:261-276.

3. Colcher, D., et al . , Pharmacokinetics and bio- distribution of genetically-engineered antibodies . Q. J. Nucl. Med. 1998, 42:225-241.

4. Daly, T., et al., Recognition of human colon cancer by T cells transduced with a chimeric receptor gene. Cancer Gene Ther. 2000, 7:284-291.

5. Fagerberg, J., et al . , Humoral anti-idiotypic and anti-anti-idiotypic immune response in cancer patients treated with monoclonal antibody 17-1A. Cancer Immunol. Immunother. 1996, 42:81-87.

6. Gruber, R. , et al . , The human anti-mouse immunoglobulin response and the anti-idiotypic network have no influence on clinical outcome in patients with minimal residual colorectal cancer treated with monoclonal antibody C017-1A. Cancer Res. 2000, 60:1921-1926.

7. Herlyn, D., et al., Anti-idiotype and recombinant antigen in immunotherapy of colorectal cancer. Cell Biophys. 1994, 24-25:143-153. 8. Herlyn, M., et al., CO 17-1A and related monoclonal antibodies: their production and characterization. Hybrido a 1986, 5 Suppl 1-.S3-10.

9. Hoffman, D.M., et al . , Adoptive cellular therapy. Semin. Oncol. 2000, 27:221-233.

10. Kufer, P., et al . , Construction and biological activity of a recombinant bi-specific single-chain antibody designed for therapy of minimal residual colorectal cancer. Cancer Immunol. Immunother. 1997, 45:193-197.

11. Lanier, L.L., et al., Subpopulations of human natural killer cells defined by expression of the

Leu-7 (HNK-1) and Leu-11 (NKI-15) antigens. J. Immunol. 1983, 131:1789-1796.

12. Lotzova, E., et al . , Immunobiology of natural killer cells. Boca Raton, CRC Press, 1986.

13. Lum, L.G., Immunotherapy with Activated T Cells after High Dose Chemotherapy and PBSCT for Breast Cancer. Proc. of the 10th Int' 1 Symposium on Autologous Blood and Marrow Transplantation 2000; In Press .

14. Mack, M., et al., Biologic properties of a bi- specific single-chain antibody directed against 17- 1A (EpCAM) and CD3: tumor cell-dependent T cell stimulation and cytotoxic activity. J. Immunol. 1997, 158:3965-3970.

15. Mack, M., et al., A small bi-specific antibody construct expressed as a functional single-chain molecule with high tumor cell cytotoxicity. Proc. Natl. Acad. Sci. USA 1995, 92:7021-7025.

16. Moritz, D., et al., A spacer region between the single chain antibody and the CD3 zeta-chain domain of chimeric T cell receptor components is required for efficient ligand binding and signaling activity.

Gene Ther. 1995, 2:539-546.

17. Pear, W.S., et al . , Production of high-titer helper- free retroviruses by transient transfection. Proc. Natl. Acad. Sci. USA 1993, 90:8392-8396.

18. Ren-Heidenreich, L., et al . , Specific targeting of EGP-2+ tumor cells by primary lymphocytes modified with chimeric T cell receptors. Hum. Gene Ther. 2000, 11:9-19.

19. Ren-Heidenreich, L. and Lum, L., Life or death of T cells with antigen-specific receptors. Current Gene

Ther. 2001, 1:253-255.

20. Rosenberg, S.A., et al . , A new approach to the adoptive immunotherapy of cancer with tumor- infiltrating lymphocytes. Science 1986, 233:1318- 1321.

21. Sen, M., et al . , Use of anti-CD3 x anti-HER2/neu bi- specific antibody for redirecting cytotoxicity of activated T cells toward HER2/neu Tumors. Journal of Hematotherapy & Stem Cell Research 2001; 10:247-260.

22. Wang, H., et al., Bi-specific antibodies in cancer therapy. Adv. Exp . Med. Biol. 2000, 465:369-380. 3. Strnad, J., et al., Molecular cloning and characterization of a human adenocarcinoma/epithelial cell surface antigen complementary DNA. Cancer Res. 1989, 49:314-317.

24. Talac, R. and Nelson, H., Current perspectives of bi-specific antibody-based immunotherapy. J. Biol. Regul. Homeost. Agents 2000, 14:175-181.

25. Todorovska, A., et al., Design and application of diabodies, triabodies and tetrabodies for cancer targeting. J. Immunol. Methods 2001, 248:47-66.

26. van Spriel, A.B., et al., Immunotherapeutic perspective for bi-specific antibodies. Immunol. Today 2000, 21:391-397.

Claims

What is claimed is:

1. A composition of matter comprising a first antigen- binding moiety and a second antigen-binding moiety operably affixed to one another via a flexible linker moiety.

2. The composition of claim 1, wherein the flexible linker moiety comprises a polymer.

3. The composition of claim 1, wherein the flexible linker moiety comprises a polypeptide.

4. The composition of claim 3, wherein the polypeptide has a length of at least 16 amino acid residues.

5. The composition of claim 4, wherein the polypeptide has a length of between 16 amino acid residues and about 100 amino acid residues.

6. The composition of claim 5, wherein the polypeptide has a length of between 50 amino acid residues and about 75 amino acid residues.

7. The composition of claim 6, wherein the polypeptide has a length of about 63 amino acid residues .

8. The composition of claim 7, wherein the polypeptide comprises the amino acid sequence encoded by nucleotide 2170-2358 shown in Figures 20-1 to 20-15 (SEQ ID NO:l) .

9. The composition of claim 3, wherein the polypeptide comprises all or a portion of an antibody hinge region .

10. The composition of claim 1, wherein the first and second antigen-binding moieties specifically bind to different antigens .

11. The composition of claim 10, wherein the first antigen-binding moiety specifi-cally binds to a tumor cell surface antigen.

12. The composition of claim 10, wherein the first antigen-binding moiety specifically binds to a CD3+ cell surface antigen.

13. The composition of claim 10, wherein the first antigen-binding moiety specifically binds to a tumor cell surface antigen, and the second antigen-binding moiety specifically binds to a CD3+ cell surface antigen.

14. The composition of claim 13, wherein the tumor cell surface antigen is EpCAM, and the CD3+ cell surface antigen is CD3.

15. The composition of claim 14, wherein the first antigen-binding moiety comprises the antigen-binding portion of an anti-EpCAM antibody, and the second antigen-binding moiety comprises the antigen-binding portion of the antibody designated OKT3.

16. The composition of claim 15, wherein the anti-EpCAM antibody comprises the antigen-binding portion of the antibody designated GA733.2.

, 17. A polypeptide comprising the amino acid sequence set forth in Figures 20-1 to 20-15 (SEQ ID NO: 2) .

18. A polypeptide comprising the amino acid sequence set forth in Figure 25 (SEQ ID NO:4).

19. The composition of claim 1, wherein each antigen- binding moiety comprises the antigen-binding portion of an antibody.

20. The composition of claim 19, wherein each antigen- binding portion of the antibody is a Fab portion.

21. The composition of claim 19, wherein the antibody is chimeric .

22. The composition of claim 3, wherein the composition comprises a single polypeptide chain which forms the first and second antigen-binding moieties and the linker moiety.

.23. The composition of claim 22, wherein each of the first and second antigen-binding moieties further comprises a second polypeptide chain.

24. A nucleic acid encoding a polypeptide comprising a first antigen-binding moiety and a second antigen- binding moiety operably affixed to one another via a flexible linker moiety having a length of at least 16 amino residues.

25. The nucleic acid of claim 24 having the nucleotide sequence shown in- Figures 20-1 to 20-15 (SEQ ID NO:l) .

26. The nucleic acid of claim 24 having the nucleotide sequence shown in Figure 24 (SEQ ID NO: 3) .

27. The nucleic acid of claim 24, wherein the nucleic acid is DNA or RNA.

28. The nucleic acid of claim 27, wherein the nucleic acid is DNA.

29. The nucleic acid of claim 24, wherein the nucleic acid is an expression vector.

30. The nucleic acid of claim 29, wherein the expression vector is selected from the group consisting of a plasmid, a cosmid, a bacteriophage and a eukaryotic virus.

31. The nucleic acid of claim 30, wherein the eukaryotic virus is an adenovirus or a retrovirus.

32. A host-vector system comprising a host cell transfected with the expression vector of claim 29.

33. A method for producing a polypeptide comprising a first antigen-binding moiety and a second antigen- binding moiety operably affixed to one another via a flexible linker moiety having a length of at least 16 amino residues, which method comprises (a) culturing the host-vector system of claim 32 under conditions permitting the expression of the polypeptide, and (b) recovering the polypeptide so expressed.

34. A composition of matter comprising (a) the composition of claim 1 and (b) a cell having on its surface the antigen to which the first antigen- binding moiety specifically binds .

35. The composition of claim 34, wherein the cell is a CD3+ cell and the first antigen-binding moiety specifically binds to CD3.

36. The composition of claim 35, wherein the cell is a T-cell, the first antigen-binding moiety comprises the antigen-binding portion of the antibody designated OKT3, and the second antigen-binding moiety comprises the antigen-binding portion of the antibody designated GA733.2.

37. The composition of claim 34, wherein the composition of (a) is present in a ratio of from about 5-500 ng per million cells of (b) .

38. A method for increasing the activity of a CD3+ cell comprising contacting the cell with the composition of claim 1.

39. A method for treating a subject afflicted with a disorder mediated by the presence of an abnormal cell, comprising administering to the subject (a) an agent known to ameliorate the disorder via contact with the abnormal cell, and (b) the composition of claim 1, wherein the first antigen-binding moiety specifically binds to an antigen present on the agent, and the second antigen-binding moiety specifically binds to an antigen present on the abnormal cell.

40. The method of claim 39, wherein the subject is selected from the group consisting of a cow, a horse, a sheep, a pig, a dog, a cat, a rabbit and a primate .

41. The method of claim 40, wherein the subject is a human.

42. The method of claim 39, wherein the disorder is a tumor.

43. The method of claim 42, wherein the agent is a CD3+ cell, the first antigen-binding moiety specifically binds to CD3, and the second antigen-binding moiety specifically binds to EpCAM.

44. The method of claim 39, wherein the composition comprises the polypeptide whose amino acid sequence is shown in Figures 20-1 to 20-15 (SEQ ID NO : 2 ) .

45. The method of claim 39, wherein the composition comprises the polypeptide whose amino acid sequence is shown in Figure 25 (SEQ ID NO:4).

46. A method for treating a subject afflicted with a tumor comprising administering to the subject (a) Interleukin-2 (IL-2), (b) T cells, and (c) the antibody designated E3Bi .

47. The method of claim 46, wherein the T cells are activated T cells .

48. The method of claim 46, wherein the T cells are non- activated T cells.

49. The method of claim 46, wherein the subject is selected from the group consisting of a cow, a horse, a sheep, a pig, a dog, a cat, a rabbit and a primate .

50. The method of claim 49, wherein the subject is a human .

51. A kit for use in treating a subject afflicted with a disorder mediated by the presence of an abnormal cell, comprising (a) the composition of claim 1, wherein the first antigen-binding moiety specifically binds to an antigen present on an agent known to ameliorate the disorder and the second antigen-binding moiety specifically binds to an antigen present on the abnormal cell, and ^' (b) instructions for use.

52. A kit for use in treating a subject afflicted with a disorder mediated by the presence of an abnormal cell, comprising (a) the composition of claim 1, and (b) the agent known to ameliorate the disorder.

53. The kit of claim 51 or 52, wherein the composition of (a) comprises a polypeptide having the sequence shown in Figures 20-1 to 20-15 (SEQ ID NO: 2) .

54. The kit of claim 51 or 52, wherein the composition of (a) comprises a polypeptide having the sequence shown in Figure 25- (SEQ ID NO : 4 ) .

55. A kit for use in treating a subject afflicted with a tumo comprising (a) Interleukin-2 (IL-2) , (b) T cells, (c) the antibody designated E3Bi, and (d) instructions for use.

56. The kit of claim 55, wherein the T cells are activated T cells.

7. The kit of claim 55, wherein the T cells are non- activated T cells.