WO2018140845A2 - Bi-specific antibodies to cd64 and a disease antigen - Google Patents

Abstract

An antibody-based immunotherapeutic has application for the treatment of GBM and other common neoplasms and diseases. The exquisite epitope-binding-specificity imparted by monoclonal antibodies (mAbs) provides an ideal platform for precisely targeted immunotherapy. Using fully human or humanized mAbs drastically reduces the risk of immunogenicity against the drug and increases clinical safety. Complications associated with murine antibodies previously used in the clinic can be entirely averted. These include cytokine release syndrome^24,25 and human anti-mouse antibody (HAMA) formation leading to rapid clearance from patients' serum;²⁶ unpredictable dose-response relationships;^23,24 and an acute, potentially severe influenza-like syndrome^23,24,27,28.

Description

BI-SPECIFIC ANTIBODIES TO CD64 AND

A DISEASE ANTIGEN

RELATED APPLICATION DATA

[01] This application claims the benefit of U.S. Patent Application Ser. No. 62/451,282, filed January 27, 2017. The contents of that patent application are expressly incorporated herein.

STATEMENT OF GOVERNMENT INTERESTS

[02] This invention was made with government support under NS085412 awarded by the National Cancer Institute. The government has certain rights in the invention.

TECHNICAL FIELD OF THE INVENTION

[03] This invention is related to the area of antibody constructs. In particular, it relates to bispecific activators of myeloid cells.

BACKGROUND OF THE INVENTION

[04] Glioblastoma multiforme (GBM) remains uniformly lethal despite aggressive, image- guided tumor resection;¹ high-dose external beam radiotherapy² or brachytherapy;³ optimized chemotherapy;⁴ and recent advances in anti -angiogenic treatments.⁵ It is also the most common of the malignant primary brain tumors, the most frequent cause of cancer death in children and young adults.⁶ Moreover, current therapy is incapacitating⁷ as a result of non-specific, dose-limiting toxicity.

[05] In contrast, immunotherapy promises an exquisitely precise approach. Microglia and macrophages are the predominant immune cell infiltrating gliomas,⁸ providing an attractive immune effector cell type for therapeutic targeting. [06] In addition to being expressed on microglia and macrophages, the CD64 receptor is constitutively expressed on other myeloid cells, such as monocytes¹¹ and monocyte- derived dendritic cells (mo-DCs)²⁰, and is inducible on neutrophils via interferon gamma (INF-γ) or granulocyte colony-stimulating factor (G-CSF) stimulation,¹¹ providing a large repertoire of effector cells capable of being redirected with our technology. Specifically, aggregation of the CD64 receptor on these cells leads to phagocytosis, superoxide generation, mediator release [including tumor necrosis factor alpha (TNF- a), interleukin-1 (TL-1), and interleukin-6 (TL-6)], enhanced antigen presentation, and tumor antibody-dependent cell-mediated cytotoxicity (ADCC).¹¹

[07] EGFRvin is a constitutively activated tyrosine kinase that enhances cell growth and migration^{12 13} and confers radiation¹⁴ and chemotherapeutic^{15 16} resistance. EGFRvIII is also transferred to other cells in microvesicles¹⁷ and exerts a paracrine oncogenic effect.¹⁷ Importantly, EGFRvin is completely absent from normal tissues but frequently expressed on the surface of GBM¹⁸ and many other common neoplasms, making it an ideal immunotherapy target.¹⁹

[08] There is a continuing need in the art to develop new therapeutic agents to successfully treat those cancers which have been refractory to available treatment modalities.

SUMMARY OF THE INVENTION

[09] According to one aspect of the invention a bispecific scFv polypeptide that mediates antigen-dependent cellular cytotoxicity is provided. It comprises a first single chain variable region which binds to a tumor cell surface antigen in series with a second single chain variable region which binds to CD64.

[10] According to another aspect of the invention a polynucleotide is provided which comprises a sequence encoding a bispecific scFv polypeptide that mediates antigen- dependent cellular cytotoxicity. The polypeptide comprises a first single chain variable region which binds to a tumor cell surface antigen, in series with a second single chain variable region which binds to CD64. [11] Another aspect of the invention is a method of treating a patient who has a tumor. An amount of a bispecific scFv polypeptide that mediates antigen-dependent cellular cytotoxicity is administered to the patient. The amount is sufficient to induce an immune response to the tumor in the patient. The polypeptide comprises a first single chain variable region which binds to a tumor cell surface antigen in series with a second single chain variable region which binds to CD64.

[12] According to another aspect of the invention a method of making a bispecific scFv polypeptide that mediates antigen-dependent cellular cytotoxicity is provided. A cell comprising a polynucleotide is cultured in a culture medium such that the bispecific polypeptide is expressed. The polynucleotide encodes a polypeptide that comprises a first single chain variable region which binds to a tumor cell surface antigen in series with a second single chain variable region which binds to CD64. The bispecific scFv polypeptide is collected from the cells or culture medium.

[13] Another aspect of the invention is a method of loading myeloid cells for tumor antigen presentation. Myeloid cells are incubated in vitro with a bispecific scFv polypeptide that mediates antigen-dependent cellular cytotoxicity and tumor cells, forming myeloid cells loaded with fragments of the bispecific scFv polypeptide. The polypeptide comprises a first single chain variable region which binds to a tumor cell surface antigen in series with a second single chain variable region which binds to CD64.

[14] According to one aspect of the invention a method of loading myeloid cells for antigen presentation is provided. Myeloid cells obtained from a patient are incubated in vitro with a bispecific scFv polypeptide, and cells of a pathological tissue obtained from the patient. Myeloid cells loaded with fragments of the cells of the pathological tissue are formed. The bispecific scFv polypeptide mediates antigen-dependent cellular cytotoxicity and comprises a first single chain variable region which specifically binds to a surface antigen of the cells of the pathological tissue, in series with a second single chain variable region which binds to CD64. The loaded myeloid cells are then administered to the patient to induce an immune response against the cells of the pathological tissue. [15] These and other embodiments which will be apparent to those of skill in the art upon reading the specification provide the art with reagents and methods for treating human cancers.

BRIEF DESCRIPTION OF THE DRAWINGS

[16] Figs. 1 A-1D show alternate construct design for hEGFRvIII-CD64 constructs.

[17] Fig. 2 (Table 1) shows CD64 receptor biological functions upon receptor cross-linking

[18] Fig. 3 (Table 2) shows previous CD64-targeting bispecific antibody constructs

[19] Fig. 4 shows that hEGFRvIII-CD64 bi-scFv redirects macrophages to induce specific lysis of EGFRvIII-positive glioma cells. IFN-gamma stimulated U937 cells were plated with EGFRvIII-positive glioma cells (U87-MG-EGFRvIII) at a ratio of 10: 1 in a standard chromium release assay. hEGFRvIII-CD64 bi-scFv was added at the indicated concentrations and specific lysis was measured at 20 hours.

[20] Fig. 5. Schematic representing EGFRvIII-specific Bispecific Activators of Myeloid Cells (BAM) binding to a tumor cell via the tumor specific agent EGFRvIII and a macrophage via CD64. The EGFRvIII binding portion does not bind to the wild-type EGFR.

[21] Fig. 6. Wild type and mutant EGFR are shown.

[22] Figs. 7A-7D. Fluorescence-activated cell sorting (FACS) analysis demonstrating purified hEGFRvIII-CD64 BAM binding specifically to macrophages and EGFRvIII. One xlO⁶ EGFRvIII^⁰ (Fig. 7 A) or EGFRvIII^P0S (Fig. 7B) malignant glioma cells are incubated with 1 microgram of hEGFRvIII-CD64 BAM and washed using standard protocols. Tumor binding BAM was detected using secondary antibody. One xlO⁶ thyoglycolate generated human transgenic macrophages were incubated with 1 microgram of hEGFRvIII-CD64 BAM and fluorescently labeled anti-CD l ib and F4/80 antibody and washed using standard protocols (Fig. 7C). Macrophage binding BAM was detected using secondary antibody (Fig. 7D). [23] Fig. 8. Non-reduced SDS PAGE demonstrating purity of hEGFRvIII-CD64 BAM following (1) affinity chromatography (lane 2) and (2) size exclusion chromatography (lane 3). Molecular weight markers are shown in lane 1.

Fig. 9. Human EGFRvIII-CD64 bi-scFv induces superoxide release in a tumor- antigen, target-specific fashion. U937 cells were plated with surface bound EGFRvIII peptide or irrelevant peptide. hEGFRvIII-CD64 bi-scFv was added and superoxide release was measured over time. In separate wells, anti-CD64 monoclonal antibody (mAb) was added to U937 cells as a positive control. Note the target specificity imparted by use of a monovalent bi-scFv as compared to bivalent mAbs which crosslink CD64 receptors and induce superoxide release in the absence of tumor peptide.

A sequence listing forms part of this disclosure. SEQ ID NO: 1 represents a nucleotide sequence optimized for expression in CHO cells. Due to codon degeneracy, alternate nucleotide sequences can code for the same protein. SEQ ID NOs: 1, 2, and 3 represent an optimized arrangement of fragments, however, other arrangements are possible. Other signal peptides are possible other than the one shown in SEQ ID NO: 4. Other linkers are possible other than the ones shown in SEQ ID NOs: 9 and 10.

DETAILED DESCRIPTION OF THE INVENTION

[26] Bispecific single chain variable region (scFv) antibodies can be made that are capable of mediating antigen-dependent cellular cytotoxicity. These antibodies have one specificity for the high affinity human IgG receptor FcyRI (CD64), present on microglia^{9 10} and macrophages,¹¹ and a second specificity for cells or antigens of a pathological tissue or pathogen. Such antibodies are also able to redirect highly prevalent microglia and macrophages for potent anti-disease immunotherapy. These bispecific antibodies have immunotherapeutic applications for the treatment of cancers, and other diseases, including without limitation: hypophosphatasia, Alzheimer's Disease, amyloidosis, C. difficile associated disease, H. pylori associated disease, hypercholesterolemia, lymphoma, multiple myeloma, HIV infection, multiple sclerosis, asthma hypereosinophilic syndrome, eosinophilic esophagitis, osteoarthritis, chronic pain, systemic lupus erythematosus, leukemia, osteoporosis, autoimmune diseases, angiogenic disease, inflammatory disorders, rheumatoid arthritis, chronic viral infections, allergies, cardiovascular disease, hematological disease, neurological disease, ophthalmic disease, such as wet AMD or ocular melanoma, orthopedic diseases, and other viral and bacterial infections, acute or chronic.

[27] Bispecific Activators of Myeloid Cells (BAMs) represent a class of molecules we have developed which are monomeric polypeptides consisting of two sFvs translated in tandem with a total molecular weight of approximately 55 kDa. These constructs possess one effector-binding arm specific for the Fc receptor CD64 while the opposing target-binding arm can be directed against "targets" that are expressed on the surface of tumor cells or other pathological cells or agents. This divalent design allows BAMs to create a molecular "tether" resulting in highly-localized and specific myeloid cell activation with concomitant anti-tumor functions (Fig. 1). These two arms are in series, which means that they are part of a continuous polypeptide chain, without implication of a particular order.

[28] BAMS may induce immunological synapses between either monocytes, macrophages, microglia, neutrophils, or dendritic cells and tumor cells. Given the small size of the BAM construct, those synapses may be indistinguishable in composition, size and subdomain arrangement from native synapses. Following BAM engagement, crosslinking CD64 receptors leads to a host of beneficial anti-tumor immune responses including: 1) tumor antibody-dependent cell-mediated cytotoxicity (ADCC)³⁷; 2) immune responses beneficial in reversing the immunosuppressive tumor microenvironment including enhanced free radical production and proinflammatory cytokine production and 3) phagocytosis and enhanced antigen presentation. Our human CD64 and EGFRvIII-targeted bispecific antibody (hEGFRvni-CD64 bi-scFv) is monovalent for both EGFRvin and CD64 and, unlike traditional monoclonal antibodies, which are bivalent, activates CD64 signaling only in the context of concurrent tumor cell binding, allowing for entirely tumor-cell- specific immunotherapy.

[29] As one particular example, to redirect highly prevalent microglia and macrophages for potent cancer immunotherapy, we have developed bispecific antibodies with specificity to the high affinity human IgG receptor FcyRI (CD64), present on microglia^{9 10} and macrophages,¹¹ and to tumor cells that express the tumor-specific epidermal growth factor receptor mutation, EGFRvIII. These bispecific antibodies have immunotherapeutic applications for the treatment of glioblastoma (GBM) and many other common neoplasms, including without limitation lung, breast, prostate, head and neck, skin, astrocytoma, and oligodendroglioma tumors.

[30] Our human anti-human CD64 and EGFRvIII-targeted bispecific antibodies (hEGFRvIII-CD64 bi-scFvs) are monovalent for both EGFRvIII and CD64 and, unlike traditional monoclonal antibodies, which are bivalent, result in CD64 aggregation and the modulation of immune effector cells only in the context of concurrent tumor cell binding. As downstream effects of CD64 binding occur only in the context of receptor aggregation,¹¹ our hEGFRvIII-CD64 bi-scFvs allow for entirely tumor-cell-specific immunotherapy. Our hEGFRvIII-CD64 bi-scFv binds to CD64 outside of the Fc binding region such that receptor binding and downstream signaling can occur even in the presence of receptor-saturating levels of IgG.

[31] Specifically, we have built hEGFRvIII-CD64 bi-scFvs by genetically linking nucleotides coding for two single chain variable fragments (scFvs) of different specificities with nucleotides coding for a short polypeptide linker (Fig. 1). Upon transcription and translation, a single polypeptide with affinity for both of the targets is expressed. Any human or humanized antibodies specific for EGFRvIII and CD64 may be used to make such bi-scFvs. One particular construct of hEGFRvIII-CD64 bi- scFvs are composed of the variable fragments of the fully human anti-EGFRvIII antibody clone 139²¹ and the humanized anti-CD64 antibody clone H22,²² drastically reducing the potential for immunogenicity and increasing clinical safety.²³'²⁴ Through the use of human antibody segments, murine-antibody-associated complications, including cytokine release syndrome²⁴'²⁵ and human anti-mouse antibody (HAMA) formation that leads to rapid clearance from patient serum, unpredictable dose- response relationships,^{23 24} and an acute, potentially severe influenza-like syndrome²³'²⁴'^{27 28} are entirely averted. We have tested two of the four possible N- to C-terminal arrangements of the variable heavy chain and light chain segments to obtain a bi-scFv construct with favorable target-binding kinetics (see Fig. 1).

[32] To express hEGFRvIII-CD64 bi-scFvs, we developed Chinese Hamster Ovary (CHO) cell codon-optimized polynucleotide sequences for each of the constructs and cloned each of these sequences into mammalian expression vectors. These vectors were transiently transfected into suspension-adapted CHO cells grown in a chemically defined, serum-free medium. Inclusion of nucleotides coding for a signal peptide at the 5' end of our constructs allows the proteins to be secreted from cultured cells (Fig. 1). Codon optimization of the polynucleotide sequences allows for robust expression in CHO cells. The ease of expression in this commonly used system will allow for straightforward integration with existing commercial manufacturing infrastructures. We are currently working to produce a stable cell line genetically engineered to express high titers of our construct. Other cell lines of other species can be used as alternatives.

[33] We have purified the protein constructs from cell culture supernatant using either genetically encoded affinity tags (e.g., poly-histidine) or a tag-free purification system that we have developed. Our tag-free purification system is based on affinity to the polypeptide spanning the EGFRvIII binding epitope (ΡΕΡνΙΠ). We have covalently conjugated the ΡΕΡνΠΙ peptide (via the terminal cysteine residue and haloacetyl chemistry) to highly cross-linked 4% agarose beads commonly used for column affinity chromatography and larger-scale industrial purification systems, allowing for affinity-based purification without extraneous purification tags. Subsequent size exclusion chromatography has allowed us to obtain highly pure, homogenous, monomelic protein.

[34] The recombinant bi-scFv constructs have proven superior to other bispecific antibody constructs, such as chemically linked Fab' fragments, which have led to the recent FDA approval of a CD3-CD19-targeted bi-scFv (blinatumomab). The advantages of using a bi-scFv construct may include, but are not limited to: 1) a smaller final protein size, allowing for better tumor penetrance; 2) eliminating the need for complex chemical cross-linking allowing for better drug homogeneity and straightforward manufacturing and regulatory considerations; and 3) the ability to bring target cells closer together which has been shown to result in more effective downstream receptor signaling as exemplified by CD3-CD19 bi-scFvs.²⁹ Other bispecific antibodies targeting CD64 are listed below in Table 2, but these are not bi-scFvs.

[35] The hEGFRvIII-CD64 bi-scFvs may be used to redirect the highly prevalent CD64- positive tumor-infiltrating immune effector cells against EGFRvIII-positive glioma, providing a viable adjunct to current standard-of-care therapy for malignant glioma and other types of cancer. The hEGFRvIII-CD64 bi-scFvs, like other BAMs, can be administered to patients alone or in conjunction with G-CSF, INF-γ, or other molecules that stimulate macrophages, neutrophils, or other immune effector cells to differentiate to effective immune cell phenotypes. Additionally, our hEGFRvIII-CD64 bi-scFvs may be used ex vivo to load antigen-presenting cells (APCs) with tumor antigens and stimulate APCs to enhance antigen presentation. These tumor antigen- loaded APCs may then be used in downstream applications to mount an effective anticancer immune response. This approach may yield highly efficacious personalized cancer vaccines. Each of these two therapeutic approaches may overcome critical limitations to the current standard-of-care for GBM and other common neoplasms.

[36] While antibodies are present in the central nervous system (CNS) in physiologic states,⁴⁵ glioma-induced changes render lesions particularly susceptible to antibody- based immunotherapy. Glioma tumor cells induce compositional changes in the basal lamina and astrocytic components of the neurovascular unit (NVU), disrupting the integrity of the blood-brain barrier (BBB). In addition to increasing tumor burden and heightening tumor invasion of the surrounding parenchyma,⁴⁶ this allows for enhanced penetrance of large soluble molecules, such as antibodies, from the vascular compartment. For the treatment of GBM, several studies have demonstrated that intravenously (IV) administered antibodies gain access to intracranial (IC) tumors and exert significant therapeutic benefit.^47"50 In murine GBM models, the anti-tenascin antibody (81C6) directed against a component of the tumor stroma showed significant localization and therapeutic activity following systemic administration, ' and, in clinical trials, IV administration of radiolabeled 81C6 showed selective tumor localization.⁵⁰ The antibody also accumulated in other tissues expressing high levels of tenascin, including the spleen, bone marrow, and liver.

[37] The vast majority of proteins found on the surface of tumor cells are also expressed on normal healthy tissue. While overexpression of specific surface antigens is characteristic of various tumors, often these antigens are tumor-associated antigens that are also expressed on the surface of healthy cells. Targeting such tumor- associated antigens via immunotherapeutic methods may risk induction of autoimmunity and thereby undermine the specificity imparted by immunotherapeutic approaches. Tumor-specific antigens, however, occur as a result of mutations in somatic genes and, when targeted therapeutically, are far less likely to be associated with autoimmunity. Accordingly, we have developed tumor-cell-specific, antibody- based immunotherapeutics that may provide a valuable adjunct to current standard-of- care therapy for malignant glioma and other types of cancer.

[38] EGFRvin is a frequent and consistent tumor-specific mutation seen in patients with GBM, and also in breast and lung carcinoma. The mutation consists of an in-frame deletion of 801 base pairs in the extracellular portion of the wild-type receptor that generates a novel glycine residue at the fusion junction (Fig. 2). This produces a highly immunogenic, cell-surface, tumor-specific epitope. Importantly, antibodies directed against EGFRvIII do not cross react with the wild-type receptor.

[39] EGFRvin encodes a constitutively active tyrosine kinase that enhances tumor cell growth and invasion while conferring radiation and chemo-therapeutic resistance. Among patients with GBM, expression of EGFRvin is an independent negative prognostic indicator. EGFRvHI also enhances the growth of neighboring EGFRvM¹^⁰ tumor cells via cytokine-mediated paracrine signaling and through the transferring of a functionally active oncogenic receptor to EGFRvffi¹^⁰ cells through the release of lipid-raft related microvesicles. Recent research has also found that EGFRvHI is expressed in glioma stem cells (GSC), an important consideration given the paradigm that tumor stem cells (TSCs) represent a subpopulation of cells that give rise to all cells in a differentiated tumor. Besides the fact that its expression is limited to tumor tissues, EGFRvin is optimal as a BAM target given its short length, making it an epitope that is close to the cell membrane; previous studies have demonstrated that both small antigen size and minimal distance from the membrane represent critical determinants for bispecific scFv-mediated potency.

[40] The disease antigen-specific arm of the bispecific polypeptide may be any of a variety of categories. Among tumor antigens, the tumor antigen may be formed by somatic mutation, such as alpha-actinin-4; ARTC1; BCR-ABL fusion protein (b3a2); B-RAF; CASP-5; CASP-8; beta-catenin; Cdc27; CDK4; CDK12; CDKN2A; CLPP; COA-1; CS K1A1 ; dek-can fusion protein ; EFTUD2; Elongation factor 2; ETV6-AML1 fusion protein; FLT3-ITD; F DC3B; FN1; GAS7; GP MB; HAUS3; HSDL1; LDLR-fucosyltransferaseAS fusion protein; HLA-A2d; HLA-Al ld; hsp70-2; MART2; MATN; ME1; MUM- If; MUM-2; MUM-3; neo-PAP; Myosin class I; FYC; OGT; OS-9; p53; pml-RARalpha fusion protein; PPP1R3B; PRDX5; PTPRK; K-ras; N-ras; RBAF600; SIRT2; SNRPD1; SYT-SSX1 or -SSX2 fusion protein; TGF-betaRII; or Triosephosphate isomerase. It may also be an antigen that is expressed more highly in a tumor than in a normal tissue, such as adipophilin; AFM- 2; ALDHIAI; BCLX (L); BF G-4; CALCA; CD45; CD274; CPSF; cyclin Dl; DKK1; ENAH (hMena); EpCAM ; EphA3 ; EZH2; FGF5; glypican-3; G250 / MN / CArX; HER-2 / neu; ELLA-DOB; Hepsin; IDOl; IGF2B3; IL13Ralpha2; Intestinal carboxyl esterase; alpha-foetoprotein; Kallikrein 4; KTF20A; Lengsin; M- CSF; MCSP; mdm-2; Meloe; or Midkine. Alternatively, the tumor antigen can be a differentiation antigen, such as CEA ; gplOO / Pmell7; mammaglobin-A; Melan-A / MART-1; NY-BR-1; OA1; PAP; PSA; RAB38 / NY-MEL- 1; TRP-l/gp75; TRP-2; or tyrosinase. The antigen may also be one that is a shared antigen such as BAGE-1; D393-CD20n; Cyclin-Al; GAGE-1,2,8; GAGE-3,4,5,6,7; GnTVf; HERV-K-MEL; KK-LC-1; KM-HN-1; LAGE-1; LY6K; or MAGE- Al .

[41] Other disease-related antigens may also be used as targets for the bi-specific polypeptides. These may be ligands, cell surface receptors, growth factors, bacterial or viral antigens, or other cell surface or soluble targets. Specific examples of such target antigens include, without limitation, alkaline phosphatase, amyloid beta protein, amyloid protein A, C. difficile enterotoxin A or B, calcitonin gen-related peptide, CCR4, cell adhsionmolecules, cyclic ADP ribose hydrolase, dabigatran, folate receptor alpha, immune cell glycoprotein, HER3, IL-2 receptor, IL-4 receptor, IL-5, IL-5 alpha receptor, IL-13, mucin5AC, nerve growth factor, PCSK-9, PD-1L, phosphatidyl-serine, SALl, TNF-alpha, CD20, HER2, VEFG-A, Tissue Factor, and sclerostin.

[42] The bispecific polypeptides may be made with, for example, any tumor specific antibody as a component in tandem with an anti-CD64 antibody, preferably humanized antibody H22. Antibodies for other tumor-specific or tumor-associated targets that may be used include without limitation: Her-2/neu, GRP-R, GD2, CD19/CD37, CD30, c-erB-2, EGFR, IDH1, IDH2. The bispecific polypeptides are preferably made recombinantly as a single polypeptide, rather than through chemical conjugation. Preferably the antibody components will be human or humanized antibodies in order to minimize immunogenicity, cytokine release syndrome, and formation of human anti -mouse antibody.

[43] While the order of the components of a bispecific scFv polypeptide can have a strong effect on the function of the polypeptide, nonetheless, any of the arrangements of subparts can be used. See Fig. 1 showing LH HL; HL HL; LH LH; and HL LH. Linkers may be used optionally between the subparts, i.e., between the VH and VL domains and between the two monovalent binding domains for the two targets of the bispecific polypeptide. Any spacer may be used, but typically used are spacers that are (G₄S)_n, wherein n is 1-5.

[44] In order to facilitate manufacture of the polypeptide, the polynucleotide encoding it may be codon optimized for a desired production host cell. While we have used CHO cells and codon optimized for them, other host cells can be used and the polynucleotide readily codon-optimized for them. The codon optimization does not change the product— the polypeptide retains the same primary sequence.

[45] The polypeptides may be administered to a patient according to well-known techniques for administering therapeutic antibodies to patients. These are typically administered intravenously, intramuscularly, or subcutaneously. It may be desirable to administer the polypeptides in a combination regimen with other drugs which will enhance the therapy. For example, it may be desirable to administer a cytokine such as G-CSF or IFN-γ which induce cells of the immune system to differentiate. For example, these two cytokines induce CD64 on neutrophils.

[46] The polypeptides of the present invention are able to mediate ADCC. This is an mechanism important for successful treatment of cancers. Other mechanisms, such as phagocytosis, are not sufficient to mount an effective anti-tumor treatment.

[47] Brain tumors which may be treated include without limitation, anaplastic astrocytoma, medulloblastoma, metastatic lung cancer to the brain, and oligodendrogliomas. Other tumors which may be treated, without limitation include, lung cancer, lymphoma, ovarian cancer, cervical cancer, fallopian tubal cancer, kidney cancer, pancreatic cancer, non-small cell lung cancer, acute lymphoblastic leukemia, colorectal cancer, and breast cancer.

[48] Myeloid cells may be treated in vitro with the bispecific polypeptide to create a personalized medicine. The myeloid cells may be obtained from the patient, incubated in vitro with the BAM and a source of disease-related antigens (such as, e.g., tumor cells), and then re-infused to the patient. Examples of myeloid cells which may be used are monocytes, macrophages, neutrophils, basophils, eosinophils, erythrocytes, dendritic cells, and megakaryocytes or platelets. Incubations may be from minutes to hours or days. Suitable incubation times may depend on the type of the cell and how they have been obtained. Incubations may be from, e.g., 10 min. to 60 min.; 30 min. to 2 hours; 1 hour to 10 hours; 8 hours to 24 hours; 1 day to 3 days; without limitation. The myeloid cells may be from the patient to be treated. The disease-related antigens may also be obtained from the patient to be treated. Such personalization will decrease the risk of rejection, and increase the chance that relevant antigens are presented on the myeloid cells.

[49] One can perform a reaction where a patient's tumor cells are incubated with her own autologous APCs and CD64-EGFRvin bispecific antibody along with cytokines such as M-CSF, IFN-γ, GM-CSF, or others known to enhance phagocytosis, antigen presentation, and myeloid cell function. Those APCs can then be screened to see which of the patient's tumor antigens the cells are presenting. This can be a useful test to predict/select which tumor antigens are likely to be presented and therefore more likely to induce a therapeutic response. This can be used to predict beneficial antigens and guide/select future therapy against antigens other than those originally targeted.

[50] In another example, the reaction can be performed as described above. The "loaded APCs" can then be used in vitro to stimulate a patient's autologous immune effector cells {e.g., T cells or other effector cells). One can screen those stimulated immune effector cells to see which subsets have expanded and predict which novel tumor antigens are likely to be immunogenic and effective therapeutic targets. Those stimulated immune effector cells can also then be reinfused into a the patient to induce therapeutic responses.

[51] "Loaded APCs" may themselves be administered to a patient where they would induce a secondary immune response against tumor antigens other than those originally targeted. This would occur as the infused loaded APCs migrate to the lymph node, present antigens, and stimulate a secondary immune response against other targets that derive from the patient's tumor.

[52] One can use the bispecific CD64-EGFRvin antibody (or other BAMs) to "load APCs" in vivo. In this case, a patient's tumor fragments are phagocytosed by resident myeloid cells stimulated by the BAM. Loaded APCs then migrate to the lymph node where they can stimulate other immune effector cells, leading to a secondary immune response against antigens other than those initially targeted, thus amplifying the therapeutic response. Cytokines, such as M-CSF, IFN-γ and GM-CSF, known to enhance phagocytosis, APC migration, antigen presentation, and formation of secondary immune responses can be administered to enhance the process and improve therapeutic outcomes.

[53] Purification of the BAM polypeptide may employ any scheme that successfully separates the polypeptide from other polypeptide in the producing cell. While affinity purification and size exclusion chromatography were very successfully used by us, other methods which are convenient may also be used.

The above disclosure generally describes the present invention. All references disclosed herein are expressly incorporated by reference. A more complete understanding can be obtained by reference to the following specific examples which are provided herein for purposes of illustration only, and are not intended to limit the scope of the invention.

References

The disclosure of each reference cited is expressly incorporated herein.

1. Kelly, P.J. Stereotactic resection and its limitations in glial neoplasms. Stereotact Funct Neurosurg 59, 84-91 (1992).

2. Walker, M.D., et al. Randomized comparisons of radiotherapy and nitrosoureas for the treatment of malignant glioma after surgery. The New England journal of medicine 303, 1323-1329 (1980).

3. Reardon, D.A., et al. Salvage radioimmunotherapy with murine iodine-131 -labeled antitenascin monoclonal antibody 81C6 for patients with recurrent primary and metastatic malignant brain tumors: phase Π study results. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 24, 115-122 (2006).

4. Stupp, R., et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. The New Engl and journal of medicine 352, 987-996 (2005).

5. Vredenburgh, J. J., et al. Bevacizumab plus irinotecan in recurrent glioblastoma multiforme. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 25, 4722-4729 (2007).

6. Estimated New Cancer Deaths by Sex, US, 2012. (American Cancer Society, 2012).

7. Imperato, J. P., Paleologos, N.A. & Vick, N.A. Effects of treatment on long-term survivors with malignant astrocytomas. Annals of neurology 28, 818-822 (1990).

8. Hussain, S.F., et al. The role of human glioma-infiltrating microglia/macrophages in mediating antitumor immune responses. Neuro-oncology 8, 261-279 (2006).

9. Peress, N.S., Siegelman, J., Fleit, H.B., Fanger, M.W. & Perillo, E. Monoclonal antibodies identify three IgG Fc receptors in normal human central nervous system. Clinical immunology and immunopathology 53, 268-280 (1989).

10. Linnartz, B., Wang, Y. & Neumann, H. Microglial immunoreceptor tyrosine-based activation and inhibition motif signaling in neuroinflammation. International journal of Alzheimer's disease 2010(2010).

11. van de Winkel, J.G. & Capel, P.J. Human IgG Fc receptor heterogeneity: molecular aspects and clinical implications. Immunol Today 14, 215-221 (1993).

12. Boockvar, J. A., et al. Constitutive EGFR signaling confers a motile phenotype to neural stem cells. Molecular and cellular neurosciences 24, 1116-1130 (2003). 13. Pedersen, M.W., Tkach, V., Pedersen, N., Berezin, V. & Poulsen, H.S. Expression of a naturally occurring constitutively active variant of the epidermal growth factor receptor in mouse fibroblasts increases motility. Int J Cancer 108, 643-653 (2004).

14. Lammering, G., et al. Inhibition of the type ΙΠ epidermal growth factor receptor variant mutant receptor by dominant-negative EGFR-CD533 enhances malignant glioma cell radiosensitivity. Clin Cancer Res 10, 6732-6743 (2004).

15. Montgomery, R.B., Guzman, J., O'Rourke, D.M. & Stahl, W.L. Expression of oncogenic epidermal growth factor receptor family kinases induces paclitaxel resistance and alters beta-tubulin isotype expression. The Journal of biological chemistry 275, 17358-17363 (2000).

16. Nagane, M., et al. Human glioblastoma xenografts overexpressing a tumor-specific mutant epidermal growth factor receptor sensitized to cisplatin by the AG1478 tyrosine kinase inhibitor. Journal of neurosurgery 95, 472-479 (2001).

17. Al-Nedawi, K., et al. Intercellular transfer of the oncogenic receptor EGFRvUI by microvesicles derived from tumour cells. Nature cell biology 10, 619-624 (2008).

18. Moscatello, D.K., et al. Frequent expression of a mutant epidermal growth factor receptor in multiple human tumors. Cancer Res 55, 5536-5539 (1995).

19. Wikstrand, C.J., et al. Monoclonal antibodies against EGFRvUI are tumor specific and react with breast and lung carcinomas and malignant gliomas. Cancer Res 55, 3140-3148 (1995).

20. Langlet, C, et al. CD64 expression distinguishes monocyte-derived and conventional dendritic cells and reveals their distinct role during intramuscular immunization. Journal of immunology 188, 1751-1760 (2012).

21. Weber, R. Antibodies directed to the deletion mutants of epidermal growth factor receptor and uses thereof. US 7,628,986 B2. (Amgen Fremont Inc.). United States. Dec. 8, 2009.

22. Graziano, R.F., et al. Construction and characterization of a humanized anti-gamma-Ig receptor type I (Fc gamma RI) monoclonal antibody. Journal of immunology 155, 4996-5002 (1995).

23. Carter, P.J. Potent antibody therapeutics by design. Nat Rev Immunol 6, 343-357 (2006).

24. Hansel, T.T., Kropshofer, H., Singer, T., Mitchell, J. A. & George, A.J. The safety and side effects of monoclonal antibodies. Nature reviews. Drug discovery 9, 325-338 (2010). 25. Gaston, R.S., et al. OKT3 first-dose reaction: association with T cell subsets and cytokine release. Kidney international 39, 141-148 (1991).

26. Reynolds, J.C., et al. Anti-murine antibody response to mouse monoclonal antibodies: clinical findings and implications. International journal of radiation applications and instrumentation. Part B, Nuclear medicine and biology 16, 121-125 (1989).

27. Kuus-Reichel, K., et al. Will immunogenicity limit the use, efficacy, and future development of therapeutic monoclonal antibodies? Clinical and diagnostic laboratory immunology 1, 365-372 (1994).

28. Mascelli, M.A., et al. Molecular, biologic, and pharmacokinetic properties of monoclonal antibodies: impact of these parameters on early clinical development. Journal of clinical pharmacology 47, 553-565 (2007).

29. Bluemel, C, et al. Epitope distance to the target cell membrane and antigen size determine the potency of T cell-mediated lysis by BiTE antibodies specific for a large melanoma surface antigen. Cancer immunology, immunotherapy : CII 59, 1197-1209 (2010).

30. Valone, F.H., et al. Phase Ia/Ib trial of bispecific antibody MDX-210 in patients with advanced breast or ovarian cancer that overexpresses the proto-oncogene HER-2/neu. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 13(1995).

31. Chen, J., Zhou, J.H., Mokotoff, M, Fanger, M.W. & Ball, E D. Lysis of small cell carcinoma of the lung (SCCL) cells by cytokine-activated monocytes and natural killer cells in the presence of bispecific immunoconjugates containing a gastrin-releasing peptide (GRP) analog or a GRP antagonist. Journal of hematotherapy 4(1995).

32. Michon, J., et al. In vitro killing of neuroblastoma cells by neutrophils derived from granulocyte colony-stimulating factor-treated cancer patients using an anti- disialoganglioside/anti-Fc gamma RI bispecific antibody. Blood 86(1995).

33. Michon, J., et al. In vivo targeting of human neuroblastoma xenograft by anti- GD2/anti-Fc gamma RI (CD64) bispecific antibody. European journal of cancer (Oxford, England : 1990) 31 A(l 995).

34. Ely, P., et al. Bispecific-armed, interferon gamma-primed macrophage-mediated phagocytosis of malignant non-Hodgkin's lymphoma. Blood 87(1996).

35. Wallace, P.K., et al. Production of macrophage-activated killer cells for targeting of glioblastoma cells with bispecific antibody to FcgammaRI and the epidermal growth factor receptor. Cancer immunology, immunotherapy : CII 49, 493-503 (2000). 36. Fury, M.G., Lipton, A., Smith, K.M., Winston, C.B. & Pfister, D.G. A phase-I trial of the epidermal growth factor receptor directed bispecific antibody MDX-447 without and with recombinant human granulocyte-colony stimulating factor in patients with advanced solid tumors. Cancer immunology, immunotherapy : CII 57, 155-163 (2008).

37. United States Cancer Statistics: National Program of Cancer Registries. Available from: http://apps.nccd.cdc.gov/uscs/ (2010).

38. Estimated Number of New Cancer Cases and Deaths by Sex, US. Available from: http://www.cancer.org (2014).

39. Stupp, R., et al. Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase ΙΠ study: 5-year analysis of the EORTC-NCIC trial. Lancet Oncol 10, 459-466 (2009).

40. Imperato, J. P., Paleologos, N.A. & Vick, N.A. Effects of treatment on long-term survivors with malignant astrocytomas. Annals of neurology 28, 818-822 (1990).

41. Hodi, F.S., et al. Improved survival with ipilimumab in patients with metastatic melanoma. The New England journal of medicine 363, 711-723 (2010).

42. Kantoff, P.W., et al. Sipuleucel-T immunotherapy for castration-resistant prostate cancer. The New England journal of medicine 363, 411-422 (2010).

43. Hoogenboom, H.R. Selecting and screening recombinant antibody libraries. Nature biotechnology 23, 1105-1116 (2005).

44. Lonberg, N. Fully human antibodies from transgenic mouse and phage display platforms. Current opinion in immunology 20, 450-459 (2008).

45. Male, D.K. Immunology of Brain Endothelium and the Blood-Brain Barrier, in Physiology and Pharmacology of the Blood-Brain Barrier, Vol. 103 (ed. Bradbury, M.W.B.) 397-415 (Springer, Berlin, 1992).

46. Lee, J., et al. Glioma-induced remodeling of the neurovascular unit. Brain research 1288, 125-134 (2009).

47. Bourdon, M.A., Coleman, R.E., Blasberg, R.G., Groothuis, D R. & Bigner, D.D. Monoclonal antibody localization in subcutaneous and intracranial human glioma xenografts: paired-label and imaging analysis. Anticancer research 4, 133-140 (1984).

48. Bullard, D.E., Adams, C.J., Coleman, R.E. & Bigner, D.D. In vivo imaging of intracranial human glioma xenografts comparing specific with nonspecific radiolabeled monoclonal antibodies. Journal of neurosurgery 64, 257-262 (1986). 49. Scott, A.M., et al. A phase I clinical trial with monoclonal antibody ch806 targeting transitional state and mutant epidermal growth factor receptors. Proc Natl Acad Sci U S A 104, 4071-4076 (2007).

50. Zalutsky, M.R., Moseley, R.P., Coakham, H.B., Coleman, R E. & Bigner, D.D.

Pharmacokinetics and tumor localization of 1311-labeled anti-tenascin monoclonal antibody 81C6 in patients with gliomas and other intracranial malignancies. Cancer Res 49, 2807-2813 (1989).

Claims

We claim:

1. A bispecific scFv polypeptide that mediates antigen-dependent cellular cytotoxicity, comprising:

a first single chain variable region which specifically binds to a tumor cell surface antigen; in series with

a second single chain variable region which binds to CD64.

2. The bispecific scFv polypeptide of claim 1 wherein tumor cell surface antigen is

EGFRvin.

3. The bispecific scFv polypeptide of claim 1 wherein the tumor cell surface antigen is selected from the group consisting of: alpha-actinin-4; ARTC1; BCR-ABL fusion protein (b3a2); B-RAF; CASP-5; CASP-8; beta-catenin; Cdc27; CDK4; CDK12; CDKN2A; CLPP; COA-1; CSNKlAl; dek-can fusion protein; EFTUD2; Elongation factor 2; ETV6-AML1 fusion protein; FLT3-ITD; FNDC3B; FNl; GAS7; GPNMB; HAUS3; HSDL1; LDLR-fucosyltransferaseAS fusion protein; HLA-A2d; HLA- Al ld; hsp70-2; MART2; MATN; ME1; MUM- If; MUM-2; MUM-3; neo-PAP; Myosin class I; NFYC; OGT; OS-9; p53; pml-RARalpha fusion protein; PPP1R3B; PRDX5; PTPRK; K-ras; N-ras; RBAF600; SIRT2; SNRPD1; SYT-SSX1 or -SSX2 fusion protein; TGF-betaRII; Triosephosphate isom erase; CEA; gplOO / Pmell7; mammaglobin-A; Melan-A / MART-1; NY-BR-1; OA1; PAP; PSA; RAB38 / NY-MEL- 1; TRP-l / gp75; TRP-2; tyrosinase; adipophilin; AFM-2; ALDHIAI; BCLX (L); BF G-4; CALCA; CD45; CD274; CPSF; cyclin Dl; DKK1;

ENAH (hMena); EpCAM; EphA3; EZH2; FGF5; glypican-3; G250 / MN / CAIX; HER-2 / neu; HLA-DOB; Hepsin; IDOl; IGF2B3; IL13Ralpha2; Intestinal carboxyl esterase; alpha-foetoprotein; Kallikrein 4; KIF20A; Lengsin; M-CSF; MCSP; mdm-2; Meloe; Midkine; BAGE-1; D393-CD20n; Cyclin-Al; GAGE- 1,2,8; GAGE-3,4,5,6,7; GnTVf; HERV-K-MEL; KK-LC-1; KM-HN-1; LAGE-1; LY6K; and MAGE-A1.

4. The bispecific scFv polypeptide of claim 1 wherein the first single chain human variable region is derived from human antibody mAbl39.

5. The bispecific scFv polypeptide of claim 1 wherein the second single chain variable region is derived from humanized antibody H22.

6. The bispecific scFv polypeptide of claim 1 wherein the first single chain human

variable region is derived from human antibody mAbl39, wherein the second single chain variable region is derived from humanized antibody H22, wherein the second single chain variable region comprises a VH domain followed by a VL domain in N- terminal to C-terminal order.

7. The bispecific scFv polypeptide of claim 1 wherein a spacer polypeptide segment links the first and second single chain variable regions.

8. The bispecific scFv polypeptide of claim 7 wherein the spacer polypeptide segment is (G4S)_n wherein n is 1-5.

9. The bispecific scFv polypeptide of claim 1 wherein a spacer polypeptide segment links VH and VL domains within the first single chain variable region.

10. The bispecific scFv polypeptide of claim 1 wherein a spacer polypeptide segment links VH and VL domains within the second single chain variable region.

11. The bispecific scFv polypeptide of claim 9 wherein the spacer polypeptide segment is (G4S)_n wherein n is 1-5.

12. The bispecific scFv polypeptide of claim 10 wherein the spacer polypeptide segment is (G4S)_n wherein n is 1-5.

13. The bispecific scFv polypeptide of claim 1 wherein each single chain variable region comprises a disulfide bond between the VH and the VL domain.

14. The bispecific scFv polypeptide of claim 1 wherein the second single chain variable region binds to CD64 outside of the Fey- binding region of CD64.

15. The bispecific scFv polypeptide of claim 1 which comprises segments encoded by SEQ ID NOs: 5 to 8.

16. The bispecific scFv polypeptide of claim 1 which comprises segments encoded by SEQ ID NOs: 5 to 8 and 10

17. The bispecific scFv polypeptide of claim 1 which comprises segments encoded by SEQ ID NOs: 4 to 8.

18. A polynucleotide comprising a sequence encoding the bispecific scFv polypeptide of claim 1.

19. The polynucleotide of claim 18 comprising the sequence of SEQ ID NO: 1.

20. The polynucleotide of claim 18 wherein the sequence encoding the bispecific scFv polypeptide is codon-optimized for production in a host cell.

21. A method of treating a patient who has a tumor, comprising:

administering to the patient an amount of the bispecific scFv polypeptide of claim 1 sufficient to induce an immune response to the tumor.

22. The method of claim 21 wherein the tumor cell surface antigen is EGFRvIII.

23. The method of claim 22 wherein the tumor is a brain tumor.

24. The method of claim 22 wherein the tumor is a glioblastoma multiforme.

25. The method of claim 21further comprising administering to the patient G-CSF or IFN- γ in an amount sufficient to induce CD64 on neutrophils.

26. A method of making a bispecific scFv polypeptide comprising:

culturing a cell comprising the polynucleotide of claim 18 in a culture medium such that the bispecific polypeptide is expressed; and

collecting the bispecific scFv polypeptide from the cells or culture medium.

27. The method of claim 26 wherein the tumor cell surface antigen is EGFRvIII.

28. The method of claim 27 further comprising:

subjecting the bispecific scFv polypeptide to an affinity reagent which comprises a peptide, wherein the peptide comprises amino acids upstream and amino acids downstream from a deletion junction found in EGFRvIII, wherein the amino acids upstream are joined to the amino acids downstream as in the EGFRvIII.

29. The method of claim 28 wherein the affinity reagent is a peptide coupled to a solid support.

30. The method of claim 29 wherein the solid support is an agarose bead.

31. The method of claim 29 wherein the affinity reagent is a column packing matrix.

32. The method of claim 28 further comprising:

subjecting the bispecific scFv polypeptide to size exclusion chromatography.

33. A method of loading myeloid cells for tumor antigen presentation, comprising:

incubating in vitro (a) myeloid cells with (b) the bispecific scFv polypeptide of claim 1 and (c) tumor cells, whereby myeloid cells loaded with fragments of the tumor cells are formed.

34. The method of claim 33 further comprising the step of : administering the loaded myeloid cells to a patient to induce an immune response against tumor cells.

35. The method of claim 34 wherein the tumor cells are obtained from the patient.

36. The method of claim 34 wherein the myeloid cells are obtained from the patient.

37. A method of loading myeloid cells for antigen presentation, comprising:

incubating in vitro (a) myeloid cells obtained from a patient, with (b) a bispecific scFv polypeptide, and (c) cells of a pathological tissue obtained from the patient, whereby myeloid cells loaded with fragments of the cells of the pathological tissue are formed, wherein the bispecific scFv polypeptide mediates antigen-dependent cellular cytotoxicity and comprises a first single chain variable region which specifically binds to a surface antigen of the cells of the pathological tissue; in series with a second single chain variable region which binds to CD64; and

administering the loaded myeloid cells to the patient to induce an immune response against the cells of the pathological tissue.

38. The method of claim 37 wherein the surface antigen is selected from the group

consisting of a viral antigen, receptor, growth factor, and a ligand.