Epidermal Growth Factor Receptor Antibodies

The present invention is related to a novel single chain antibody-toxin, scFv(14E1)-ETA, based on the independently isolated EGF receptor specific monoclonal antibody 14E1.

The erbB/EGF receptor related gene family encodes growth factor receptors with intrinsic tyrosine kinase activity. Four members of this family have been identified: ErbB/EGF receptor, ErbB-2, ErbB-3 and ErbB4 (reviewed in Peles and Yarden, 1993). Members of this family have been implicated in the development of a variety of human malignancies. EGF receptor gene amplification and overexpression has been observed in a high percentage of primary human carcinomas of epithelial origin including glioblastoma, cancer of the lung, breast, head and neck, and bladder, and correlates with an unfavorable prognosis for the patients (reviewed in Gullick, 1991). Increased receptor expression in tumor cells is often accompanied by increased production of TGF.alpha. (Derynck et al., 1987), which leads to receptor activation by an autocrine pathway and contributes to malignant transformation. Due to its accessibility on the cell surface, its overexpression in several types of cancer, and its involvement as a marker for an unfavorable prognosis, the EGF receptor and/or its variants (Ekstrand, A. J., Longo, N., Hamid, M. L., Olson, J. J., Liu, L., Collins, V. P., and James, C. D. Functional characterization of an EGF receptor with a truncated extracellular domain expressed in glioblastomas with EGFR gene amplification. Oncogene, 9: 2313-2320, 1994; Sugawa, N., Ekstrand, A. J., James, C. D., and Collins, V. P. Identical splicing of aberrant epidermal growth factor receptor transcripts from amplified rearranged genes in human glioblastomas. Proc. Natl. Acad. Sci. U.S.A., 87: 8602-6, 1990; Wong, A. J., Ruppert, J. M., Bigner, S. H., Grzeschik, C. H., Humphrey, P. A., Bigner, D. S., and Vogelstein, B. Structural alterations of the epidermal growth factor receptor gene in human gliomas. Proc. Natl. Acad. Sci. U.S.A., 89: 2965-9, 1992; Yamazaki, H., Ohba, Y., Tamaoki, N., and Shibuya, M. A deletion mutation within the ligand binding domain is responsible for activation of epidermal growth factor receptor gene in human brain tumors. Jpn. J. Cancer Res., 81: 773-9, 1990; Wikstrand, C. J., Hale, L. P., Batra, S. K., Hill, M. L., Humphrey, P. A., Kurpad, S. N., McLendon, R. E., Moscatello, D., Pegram, C. N., Reist, C. J., and et al Monoclonal antibodies against EGFRvIII are tumor specific and react with breast and lung carcinomas and malignant gliomas. Cancer Res., 55: 3140-3148, 1995; Moscatello, D. K., Holgado-Madruga, M., Godwin, A. K., Ramirez, G., Gunn, G., Zoltick, P. W., Biegel, J. A., Hayes, R L., and Wong, A. J. Frequent expression of a mutant epidermal growth factor receptor in multiple human tumors. Cancer Res., 55: 5536-9, 1995; Hills, D., Rowlinson-Busza, G., and Gullick, W. J. Specific targeting of a mutant, activated EGF receptor found in glioblastoma using a monoclonal antibody. Int. J. Cancer, 63: 537-543, 1995) are under intensive scrutiny as a therapeutic target for novel antitumor reagents.

Various strategies have been employed to target the EGF receptor and/or his variants for tumor therapy. Monoclonal antibodies directed towards the extracellular domain of the EGF receptor have proven effective in the inhibition of tumor cell growth. The EGF receptor specific Mab 225 competes with EGF for binding to the EGF receptor thereby blocking ligand dependent receptor activation (Fan et al., 1993). Treatment with Mab 225 inhibits the growth of EGF receptor expressing tumor cells in vitro and in animal models in vivo (Masui et al., 1984; Ennis et al., 1989), and clinical studies with Mab 225 have recently been initiated. In an attempt to achieve more potent antitumoral effects recombinant fusion proteins have been constructed which contain the enzymatic domains of Pseudomonas exotoxin A (Chaudhary et al., 1987) or diphtheria toxin (Shaw et al., 1991), and employ the natural EGF receptor ligands TGF.alpha. or EGF for targeting to receptor overexpressing tumor cells. Due to the growth factor domain such toxins are able to activate the EGF receptor (Schmidt and Wels, 1996), which might facilitate rapid uptake by tumor cells, but could also be responsible for the significant cytotoxic activity displayed on cells expressing only moderate levels of the target receptor.

As an alternative to growth factors, single-chain immunotoxin (SCIT) fusion proteins can be used for the target cell specific delivery of therapeutic effector functions. SCIT fusion proteins are hybrid molecules that contain antibody variable regions (scFv) genetically fused to a binding defective toxin, most notably diphteria toxin (DT) and Pseudomonas exotoxin A (ETA). Both DT and PE halt protein synthesis by ADP-ribosylating elongation factor 2 upon entry to the cell cytosol, resulting in cell death (Carroll and Collier, 1987). A recombinant single chain antibody-toxin consisting of a scFv domain derived from the antagonistic Mab 225 and truncated Pseudomonas exotoxin A was described recently (Wels et al., 1995).

It is an object of the invention to provide an EGF receptor specific single chain antibody comprising a Vlight and Vheavy chain which is highly specific for tumor cells expressing, especially overexpressing, the EGF receptor and/or EGF receptor variants for example EGFRvIII. In accordance with the invention, the single chain antibody or polypeptide can comprise a Vlight chain having the characteristics of monoclonal antibody 14E1, e.g., sequences 3, 4, 5 and 6, and a Vheavy chain having the characteristics of monoclonal antibody 14E1, e.g., sequences 1 and 2. The antibody can be conjugated or joined to a biologically-active component or diagnostic agent for therapeutic and diagnostic purposes. For example, the antibody can be joined to a cytotoxin, such as ETA, to achieve a highly potent cell killing activity. It is a further object of the invention to provide a pharmaceutical composition containing an effective amount of the mentioned single chain antibody as an active component and a non-toxic pharmaceutical carrier or solution.

It is another aspect of the invention to provide a test kit comprising an EGF receptor and/or EGF receptor variants specific single chain antibody as mentioned above, e.g., conjugated or joined to a biologically-active component such as ETA, or conjugated to a MRI contrast agent, a radiodiagnostic agent, or a radiotherapeutic agent. By biologically active, it is meant that the component has an effect in a biological system, either on an in vitro or in vivo system. Biologically-active components includes, e.g., cytotoxins such as ETA or diphteria toxin, cytokines such as interferons (.alpha.-, .beta.- and .gamma.), interleukins, TNF-.alpha., Rantes, MIP, hormones, estrogens, growth factors, etc. See, e.g., Siegall et al., Drug Development Research, 34:210-219, 1995 for other cytotoxins or immunotoxins. The biologically-active component can be covalently joined to the antibody, or noncovalently joined, e.g., by hydrogen or ionic bonds. If the biologically-active component is a peptide, it can be fused in-frame to the single chain polypeptide in a continous reading frame as illustrated in the example below for ETA. The single chain polypeptide can also comprise a radiodiagnostic or radiotherapeutic agent or a MRI contrast agent (for reference see for example DE-OS 37 10730 A1, European Patent Application 0512661A1, DE-OS 43 37 600 A1, DE-OS 43 13 670 A1 and DE-OS 43 37 599 A1). An agent can be an atom, a molecule, a compound, a chelate, etc., or any substance which is effective for these purposes. A radiodiagnostic agent can be, e.g., a radioactive isotope complexed with a chelate. A MRI agent comprises, e.g., a paramagnetic ion. A single chain polypeptide in accordance with the present invention also includes polypeptides which have the characteristics of monoclonal antibody 14E1. By the phrase "has the characteristics of," it is meant that the polypeptide binds to the same epitope or to a mimetope which is also specifically recognized by mAb 14E1. The polypeptide can also have substantially the same binding affinity and/or tissue specificity and/or amino acid sequence of monoclonal antibody 14E1.

The present invention also includes a double-chain polypeptide having a binding affinity for an epidermal growth factor receptor and/or its variants, said polypeptide comprising: (1) a first polypeptide segment comprising the binding portion of the heavy chain variable domain of monoclonal antibody 14E 1; and (2) a second polypeptide segment comprising the binding portion of the light chain variable domain of monoclonal antibody 14E1, where (1) and (2) are joined by other than peptide sequences, e.g., cysteine bonds. See, e.g. Reiter et al., Nature Biotechnology, 14:1239-1245, 1996. The binding portion is the part of the polypeptide segment which recognizes the desired epitope.

The present invention also relates to nucleic acid sequences which hybridize to the coding sequences which comprise the single chain polypeptide comprising the heavy chain variable domain of antibody 14E 1 and the light chain variable region domain of antibody 14E 1. As used here, "stringent conditions" means any conditions in which hybridization will occur where there is at least about 95%, preferably 97%, nucleotide complementarity between the nucleic acids. See, e.g., Sambrook et al.

The polypeptides of the present invention can be administered to patients in need of therapy. To treat the disease, the compound, or mixture, can be formulated into pharmaceutical composition comprising a pharmaceutically acceptable carrier and other excipients as apparent to the skilled worker. See, e.g., Remington's Phamaceutical Sciences, Eighteenth Edtion, Mack Publishing Company, 1990. Such composition can additionally contain effective amounts of other compounds, especially for treatment of cancer. The polypeptides can be administered intravenously, enterally, or parenterally. Generally, dosage ranges will be selected to achieve cytotoxic or cytostatic effects. The dosage will depend on the age, condition, sex, and extent of the disease in the patient. For example, dosages in the range of about 0.1 mg/kg can be administered, as single or multiple doses daily. A pharmaceutical composition can be used to treat various types of tumors, including carcinomas, gliomas, and melanomas.

The present invention can also be employed in vitro to purge a mixed population of cells having an antigen recognized by a polypeptide according to the present invention. For other aspects of the nucleic acids, polypeptides, antibody, etc., reference is made to standard textbooks of molecular biology, protein science, and immunology. See, e.g., Davis et a. (1986), Basic Methodes in Molecular Biology, Elsvir Sciences Publishing, Inc., New York; Hames et al. (1985), Nuceic Acid Hybridisation, IL Press, Molecular Cloning, Sambrook et al.; Current Protocols in Molecular Biology, Edited by F. M. Ausubel et al., John Wiley & Sons, inc. Current Protocols in Human Genetics, Edited by Nicholas C. Dracopoli et al., John Wiley & Sons, Inc.; Current Protocols in Protein Science; Edited by John E. Coligan et al., John Wiley & Sons, Inc.; Current Protocols in Immunology; Edited by John E. Coligan et al., John Wiley & Sons Inc.

Without further elaboration, it is believed that one skilled in the art can, using the preciding description, utilize the present invention to its fullest extent. The preciding preferred specific embodiments are, therefore, to be construed as merely illustrative, and not limitative of the remainder of the disclosure in any way whatsoever.

The entire disclosure of all applications, patents and publications, cited above and below are hereby incorporated by reference.

From the forgoing description, one skilled in the art can easily ascertain the essential characteristics of this invention, and without departing from the spirit and scope thereof, can make various changes and modifications of the invention to adapt it to various usages and conditions.

FIG. 1. Schematic representation of the recombinant single-chain antibody-toxins scFv(14E1)-ETA and scFv(225)-ETA, and the growth factor toxin TGF.alpha.-ETA. The bacterially expressed scFv-ETA proteins consist of the scFv domains of the monoclonal antibodies 225 or 14E1 containing the heavy (V.sub.H) and light chain (V.sub.L) variable domains fused to amino acids 252 to 613 of Pseudomonas exotoxin A (ETA) representing the translocation domain II, domain Ib, and domain III which mediates the ADP ribosylation of the eukaryotic elongation factor 2. TGF.alpha.-ETA contains amino acids 1 to 50 of human TGF.alpha. as an EGF receptor specific binding domain. Included in the molecules are the synthetic FLAG epitope and a cluster of 6 His residues at the N-terminus, and another cluster of 6 His residues N-terminal of ETA domain II facilitating the purification of the proteins via Ni.sup.2+ affinity chromatography (not shown).

FIG. 2. Binding of scFv(14E1)-ETA and scFv(225)-ETA to recombinant extracellular domain of the EGF receptor. Immobilized extracellular domain of the EGF receptor was incubated with various concentrations of scFv(14E1)-ETA (closed circles) or scFv(225)-ETA (open circles). The amount of specifically bound protein was measured, after incubation with rabbit anti-ETA serum followed by alkaline phosphatase coupled anti-rabbit IgG and conversion of the phosphatase substrate p-nitrophenylphosphate as the absorbance at 405 nm.

FIG. 3. Competition of the cytotoxic activity of scFv(225)-ETA and scFv(14E1)-ETA by the monoclonal antibody 225. A431 human squamous cell carcinoma cells were incubated for 40 h with 100 ng/ml scFv(225)-ETA (left) or scFv(14E1)-ETA (right) without the addition of competitor or in the presence of a 100-fold molar excess of the EGF receptor specific Mab 225 as a specific competitor, or the isotype-matched control antibody FRP5 as indicated. The relative number of viable cells was determined using an enzymatic assay as described in "background of examples". Each point represents the mean of a set of data determined in triplicate.

FIG. 4. In vitro cell killing activity of scFv-ETA proteins and TGFcx-ETA. A431 human squamous cell carcinoma cells (upper left panel), and MDA-MB468 (upper right panel) and SKBR3 (lower left panel) human breast carcinoma cells were incubated for 40 h with the indicated concentrations of scFv(225)-ETA (open circles), scFv(14E1)-ETA (closed circles), or TGF.alpha.-ETA (open boxes). In addition SKBR3 cells were treated with the ErbB-2 specific scFv(FRP5)-ETA (closed triangles). The relative number of viable cells was determined as described in FIG. 3 and "background of examples". Each point represents the mean of a set of data determined in triplicate in 3 independent experiments.

FIG. 5. Kinetics of scFv-ETA cell killing on A431 cells. The cells were incubated with 100 ng/ml of scFv(225)-ETA (closed circles), scFv(14E1)-ETA (open circles), or scFv(FRP5)-ETA (closed triangles) for various time intervals as indicated. After treatment the toxins were removed and the cells were incubated for another 40 h in normal growth medium. The relative number of viable cells was determined as described in FIG. 3 and "background of examples". Each point represents the mean of a set of data determined in triplicate.

FIG. 6. Competition of the cytotoxic activity of TGF.alpha.-ETA by scFv(225) and scFv(14E1). MDA-MB468 (upper panel), A431 (middle panel), and SKBR3 cells (lower panel) were incubated for 40 h with 100 ng/ml TGF.alpha.-ETA in the absence of competitor, or in the presence of 10 .mu.g/ml of scFv(225) or scFv(14E1) as indicated. The relative number of viable cells was determined as described in FIG. 3 and "background of examples". Each point was determined in triplicate. The standard deviation is represented by error bars.

FIG. 7. Inhibition of the in vitro growth of NIH3T3 transfectants by recombinant toxins. DH-E.DELTA.801-P cells expressing variant EGFRvIII, DH-EGFR-E cells expressing full-length EGFR, and DH-L-2 control cells were incubated for 40 h with the indicated concentrations of the EGFR specific toxins scFv(14E1)-ETA, TGF.alpha.-ETA, or with the ErbB2 specific scFv(FRP5)-ETA as a control. The relative number of viable cells in comparison to PBS treated cells was determined using an enzymatic assay described in "background of examples". Each point represents the mean of a set of data determined in triplicates in three independent experiments.

FIG. 8. Inhibition of the in vitro growth of DH-E.DELTA.801-5 cells by recombinant toxins. DH-E.DELTA.801-5 cells were incubated for 40 h with the indicated concentrations of the EGFR specific toxins scFv(14E1)-ETA or TGF.alpha.-ETA and the relative number of viable cells was determined as described in the legend of FIG. 2. For comparison a corresponding experiment with DH-EGFR-E cells is shown (data are from FIG. 7).

FIG. 9. A. Nucleotide sequence of VH 14E1 (SEQ ID NO:1); B. Amino acid sequence of VH 14E1 (SEQ ID NO:2); C. Nucleotide sequence of VL 14E1 (SEQ ID NO:3); D. Amino acid sequence of VL 14E1 (VL #2) (SEQ ID NO:4); E. Nucleotide sequence of VL 14E1 (SEQ ID NO:5); F. Amino acid sequence of VL 14E1 (VL #1) (SEQ ID NO:6); G. Nucleotide sequence of scFv(14E1) (SEQ ID NO:7); and H. Amino acid sequence of scFv(14E1) (SEQ ID NO:8).

FIGS. 10(A and B). Nucleotide and amino acid sequence of scFV(14E) SEQ ID NO:7 and SEQ ID NO:8.

Cell Lines and Cell Culture. The SKBR3 and MDA-MB468 human breast tumor cell lines, the A431 human epidermoid tumor cell line and the NE1 mouse fibroblast cell line were maintained in Dulbecco's modified Eagle's medium containing 8% heat inactivated FCS. The generation of the stably transfected NIH3T3 cell lines DH-EGFR-E expressing human full-length EGFR cDNA, DH-E.DELTA.801-P and DH-E.DELTA.801-5 expressing EGFRvIII variant cDNA, and DH-L-2 control cells is described in detail by Hills et al., 1995 (Hills, D., Rowlinson-Busza, G., and Gullick, W. J. Specific targeting of a mutant, activated EGF receptor found in glioblastoma using a monoclonal antibody. Int. J. Cancer, 63: 537-543, 1995). The cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% heat inactivated fetal calf serum at 37.degree. C. in a humidified atmosphere of air containing 5% CO.sub.2.

Preparation of the 14E1 hybridoma. The hybridoma cell line 14E1 was prepared at Triton Diagnostics Inc. (Alameda, Calif.) by fusing splenocytes from mice imununized with A431 cells to the non-secreting mouse myeloma cell line SP2/0-Ag14 (both from ATCC, Rockville, Md.) according to the method of Kohler and Milstein (1975) (Nature 256:495-497 (1975)). Briefly, Balb/c mice were intraperitoneally immunized with 10.sup.7 A431 cells emulsified in an equal volume of RIBI adjuvant (RIBI Immunochem Research Inc., Hamilton, Mont.) and boosted every two weeks. Sera were collected and tested biweekly by ELISA for reactivity against A431 cell lysate and the extracellular portion of the EGF receptor purified from A431 cell conditioned media (Science, 224: 294-297 (1984)). Mice with positive titers were intraveneously boosted with antigen in PBS and sacrificed four days later. A 4:1 ratio of splenocytes to myeloma cells were fused using a 50% polyethylene glycol solution, and plated in 96 wells at a density of 2.5.times.10.sup.5 splenocytes per well in RPMI supplemented with 0.1 mM hypoxanthine and 5.8 .mu.M azaserine to select for hybrids. Supernatants were screened by ELISA for EGF receptor specificity; positive clones were isolated and put through two additional rounds of subcloning. Antibodies were raised in ascites and purified on protein G-agarose. Several EGF receptor specific clones were isolated; one of these, 14E1 (IgG), which recognizes the extracellular portion of the EGF receptor, was shown to compete with EGF and TGF.alpha. for receptor binding in radioligand binding assays (data not shown).

Construction of scFv(14E1) and scFv(14E1)-ETA. 14E1 hybridoma cell mRNA was isolated using a Quick Prep RNA purification kit (Phainacia Biotech, Brussels, Belgium). First strand cDNA synthesis was carried out according to the manufacturer's recommendations using a cDNA synthesis kit (Stratagene) with 100 ng mRNA and random primers. For amplification of the heavy (VH) and light chain (VL) variable domains the first strand cDNA served as a template in a PCR as described (Wels et al., 1992). For amplification of the VH domain, 50 pmol each of the oligonucleotides 5'-AGGTSMARCTGCAGSAGTCWGG-3' (SEQ ID NO:9) and 5'-TGAGGAGACGGTGACCGTGGTCCCTTGGCCCC-3' (SEQ ID NO:10) were used, for amplification of the VL kappa domain, 50 pmol each of the oligonucleotides 5'-GCGACCTTGCACGCGTAGACATTGAGCTCACCCAGTCTCCA-3' (SEQ ID NO:11), and 5'-CGCTACAATAGCGGCCGCTACCGTCCGTTTGATTTCCAGCTTGGTGCC-3' (SEQ ID NO:12) or 5'-CGCTACATTAGCGGCCGCTACCGTCCGTTTCAGCTCCAGCTTGGTCCC-3' (SEQ ID NO:13) were used (M=A+C, R=A+G, S=C+G, W=A+T, Y=C+T, K=G+T). Subsequently, the VH and VL PCR products were reamplified using 50 pmol each of the oligonucleotide primers 5'-TGAGGAGACGGTGACCGTGGTCCCTTGGCCCCAG-3' (SEQ ID NO:14) and 5'-ATTATAAGCTTCAGGTSMARCTGCAGSAGTCWGG-3' (SEQ ID NO:15) (VH), or 5'-TTAGATCTCTAGAAKCTCGAGYTTKGTSC-3' (SEQ ID NO:16) and 5'-GACATTCAGCTGACCCAGWCTSC-3' (SEQ ID NO:17) (VL), respectively. PCR products were digested with HindIII and BstEII (VH), or with PvuII and XbaI (VL). Mab 14E1 VH cDNA was inserted into HindIII/BstEII digested plasmid pWW152 (Wels et al., 1992) which contains a sequence encoding the 15 amino acid linker (GGGGS).sub.3 (SEQ ID NO:18). Subsequently, the 14E1 VL fragment was inserted 3' of the VH and linker sequences, resulting in the scFv(14E1) encoding plasmid pWW152-14E1. For bacterial expression the scFv(14E1) sequence was isolated as a HindIII/XbaI fragment from pWW152 and inserted into HindIII/XbaI digested plasmid pSW50 (Wels et al., 1995). For the construction of the scFv(14E1)-ETA toxin fusion the scFv(14E1) fragment was inserted into HindIII/XbaI digested plasmid pSW202 containing a truncated Pseudomonas ETA gene which lacks the original cell binding domain Ia of the toxin (Wels et al., 1995). The resulting plasmids pSW50-14E1 and pSW202-14E1 encode, respectively, the scFv(14E1) and scFv(14E1)-ETA proteins, fused at the N-terminus to the ompA signal sequence, the synthetic FLAG epitope, and a cluster of 6 His residues, under the control of an IPTG inducible tac promoter.

Expression and purification of scFv and scFv-ETA fusion proteins. Single colonies of E. coli CC118 carrying plasmids pSW50-14E1 or pSW50-225 (Wels et al., 1995) for the expression of EGF receptor specific scFv proteins, or plasmids pSW202-14E1 or pSW202-225 (Wels et al., 1995) for the expression of scFv-ETA fusions proteins were grown overnight at 37.degree. C. in LB medium supplemented with 0.6% glucose and 100 .mu.g/ml ampicillin. E. coli CC118 carrying plasmid pSW202-TGF.alpha. (Schmidt and Wels, 1996) were used for the expression of TGF.alpha.-ETA, a recombinant growth factor toxin which consists of amino acids 1 to 50 of human TGF.alpha. fused to truncated Pseudomonas ETA. The cultures were diluted 30-fold in the same medium, grown at 37.degree. C. to an OD.sub.550 of 0.7 and induced with 0.1 mM IPTG for 1 h at room temperature. Cells were harvested by centrifigation at 10,000 g for 10 min at 4.degree. C., the cell pellet from 1 l of culture was resuspended in 15 ml PBS containing 6 M guanidine hydrochloride, and lysed by sonication. Following incubation at room temperature for 30 min the lysate was clarified by centrifugation at 30,000 g for 30 min at 4.degree. C. The supernatant was diluted to 3 M guanidine hydrochloride with PBS and recombinant proteins were purified via binding to Ni.sup.2+ -saturated chelating sepharose (Pharmacia Biotech). Specifically bound proteins were eluted with 3 M guanidine hydrochloride, 250 mM iridazole in PBS. Fractions containing recombinant fusion proteins were pooled and dialysed twice against PBS, 400 mM L-arginine and PBS, respectively. Typical yield of purified proteins was 1 mg per l of original bacterial culture with a purity of approximately 70% determined by SDS-PAGE and Coomassie brilliant blue staining.

ScFv-ETA binding assay. The binding of scFv-ETA proteins to the EGF receptor was measured by ELISA as described (Schmidt et al., 1996). 96 well microtiter plates coated with recombinant protein comprising the extracellular domain of the human EGF receptor were blocked with 3% BSA in TBS (10 mM Tris-HCl, pH 7.5, 150 mM NaCl). 50 .mu.l of scFv(225)-ETA or scFv(14E1)-ETA at concentrations ranging from 0.03 nM to 1 .mu.M were added to the wells and the plates were incubated for 1 h at 37.degree. C. Unbound protein was removed, the wells were washed and incubated with 100 .mu.l of rabbit anti-exotoxin A serum for 1 h at 37.degree. C. followed by incubation with 100 .mu.l of goat anti-rabbit IgG coupled to alkaline phosphatase (Sigma, St. Louis, Mo.). Specifically bound scFv-ETA proteins were detected by incubation with a solution of 1 M Tris-HCl, pH 8.0, 1 mg/ml p-nitrophenylphosphate disodium (Sigma) for 30 min at room temperature, then the absorbance at 405 nm was measured.

Cell viability assay. The cell killing activity of ETA fusion proteins was measured basically as described (Wels et al., 1992). The cells were seeded in 96 well plates at a density of 1.times.10.sup.4 cells/well in normal growth medium. Various concentrations of ETA fusion proteins were added to triplicate samples and the cells were incubated for 40 h. 10 .mu.l of 10 mg/ml MTT (3-(4,5-dimethylthiazole-2-yl)-2,5 diphenyltetrazolium bromide) (Sigma) in PBS were added to each well and the cells were incubated for another 3 h. Cells were lysed by the addition of 90 .mu.l of 20% SDS in 50% dimethyl formamide, pH 4.7. After solubilization of the formazan product, the OD at 590 nm of each sample was determined in a microplate reader (Dynatech) as a measure of the relative amount of viable cells in comparison to cells grown without the addition of recombinant proteins.

Competition elperiments. Competition of TGF.alpha.-ETA binding by scFv proteins was analyzed in a cell viability assay as described above. SKBR3, MDA-MB468, and A431 cells were incubated with 100 ng/ml of TGF.alpha.-ETA alone or in the presence of 10 .mu.g/ml scFv(14E1) or scFv(225) as competitors. After 40 h cell viability was determined. Competition of scFv-ETA binding by the Mab 225 was analyzed using A431 cells. The cells were incubated with 100 ng/ml of scFv(225)-ETA or scFv(14E1)-ETA with or without the addition of a 100-fold molar excess of the EGF receptor specific Mab 225 as a competitor. After 40 h the relative number of viable cells was determined as described above.

Time course of scFv-ETA cell killing. A431 cells were seeded in 96 well plates at a density of 1.times.10.sup.4 cells per well. After attatchment of the cells the medium was removed and the cells were treated for various time intervals with 100 ng/ml of scFv-ETA proteins in normal growth medium. The medium was removed, cells were washed twice with PBS, and grown for another 40 h in normal growth medium. Cell viability was measured as described above.

Receptor Activation Assay. A431 human epidermoid carcinoma cells and NE1 mouse fibroblasts expressing a human EGF receptor cDNA (Beerli et al., 1994) were grown for 16 h in DMEM supplemented with 0.5% FCS. A431 cells were treated with purified recombinant TGF.alpha.-ETA at a concentration of 1 .mu.g/ml with or without the addition of 50 .mu.g/ml scFv(14E1)-ETA or scFv(225)-ETA as competitors. Control cells were treated with PBS or 20 ng/ml EGF. NE1 cells were treated with EGF at a concentration of 10 ng/ml alone or in the presence of 20 .mu.g/ml scFv(14E1)-ETA or scFv(225)-ETA. Control cells were incubated with PBS or with scFv-ETA proteins in the absence of EGF. Following incubation at 37.degree. C. for 10 min the cells were lysed in a buffer containing 50 mM Tris-HCl pH 8.0, 5 mM EGTA, 150 mM NaCl, 1 mM PMSF, 2 mM NaVO.sub.4, 50 mM NaF, 50 Na-Molybdate, 1% Triton X-100, 0.5% desoxycholate, 0.1% SDS. Extracts were clarified by centrifuigation at 10,000 g for 10 min at 4.degree. C. Cleared cell lysates containing 15 .mu.g each of total proteins were applied on a 7.5% SDS-PAGE. After electrophoresis proteins were blotted on a PVDF membrane (Millipore) and phosphotyrosine containing proteins were detected by incubation of the membrane with an anti-phosphotyrosine Mab (Santa Cruz Biotechnology, Santa Cruz, Calif., U.S.A.), followed by incubation with an anti-mouse horse radish peroxidase coupled antibody and chermiluminescent detection with the ECL kit (Amersham).

Construction and bacterial expression of scFv(14E1)-ETA. The 14E1 hybridoma producing a novel anti-EGF receptor monoclonal antibody (IgG, .kappa.) was derived by immunization of mice with human A431 epidernoid carcinoma cells and fusion of splenocytes following standard protocols. cDNAs encoding the heavy (VH) and light chain (VL) variable domains of the 14E1 Mab were derived from 14E1 hybridoma cell mRNA by reverse transcription and amplification using PCR. A single chain Fv gene was created by connecting VH and VL sequences via a synthetic linker encoding the 15 amino acids (GGGGS).sub.3, and the scFv gene was fused to sequences encoding a truncated form of Pseudomonas aeruginosa exotoxin A (ETA) in the pSW202 vector as described (Wels et al., 1992 a; Wels et al., 1995). The resulting expression plasmid pSW202-14E1 encodes under the control of an IPTG inducible tac promoter a fusion protein consisting of the E. coli ompA signal peptide at the N-terminus, followed by the synthetic FLAG epitope, 6 His residues, the scFv(14E1), 6 His residues, and ETA amino acids 252 to 613. The structure of the scFv(14E1)-ETA gene product is schematically shown in FIG. 1A. The ETA portion of the molecule lacks the native cell binding domain Ia of the toxin but contains the translocation domain II which is required for processing of the toxin and release into the cytoplasm after interalization into target cells via the endosomal route, and the enzymatic domain III which catalyzes the ADP-ribosylation of eukaryotic elongation factor EF-2 thereby arresting cellular protein synthesis (Ogata et al., 1992).

The scFv(14E1)-ETA antibody-toxin and two previously described recombinant toxins with specificity for the EGF receptor, scFv(225)-ETA (Wels et al., 1995) and TGF.alpha.-ETA (Schmidt et al., 1996) were expressed in E. coli strain CC118. Total bacterial lysates were prepared in 6 M guanidine hydrochloride, the lysates were diluted to 3 M guanidine hydrochloride and the recombinant toxins were purified by binding to Ni.sup.+2 -saturated chelating sepharose and elution with 250 mM iridazole. Fractions containing the recombinant scFv-ETA proteins were pooled, imidazole and denaturant were removed by dialysis and the proteins were concentrated by ultrafiltration. SDS-PAGE analysis of the purified material revealed a purity of greater than 70% after a single round of Ni.sup.2+ affinity purification (data not shown). Likewise, scFv(14E1) and scFv(225) proteins which lack the C-temiinal toxin domain were expressed and purified (data not shown). The yield of purified recombinant proteins from 1 l of bacterial culture was typically 1, 0.5, and 2 mg of scFv-ETA, TGF.alpha.-ETA, and scFv proteins, respectively.

Binding properties of scFv(14E1)-ETA. ELISA experiments were performed to determine the binding of scFv(14E1) to the EGF receptor. ScFv(14E1)-ETA at concentrations ranging from 0.03 nM to 1 .mu.M was added to the wells of 96 well plates coated with purified recombinant extracellular domain of the EGF receptor, the plates were incubated at 37.degree. C. for 1 h and specifically bound protein was determined. The similar scFv(225)-ETA molecule was used as a control. The results are shown in FIG. 2A Both proteins, scFv(14E1)-ETA and scFv(225)-ETA specifically bound to the extracellular portion of the EGF receptor in a saturable fashion. The apparent binding affinity of the scFv(14E1-ETA to EGF receptor, calculated as the half-maximal saturation value, was 1 nM. The apparent binding affinity of scFv(225)-ETA to EGF receptor was much lower with a half maximal saturation value of 9 nM. Previously an apparent affinity of 12 nM was determined for scFv(225)-ETA in a similar ELISA experiment with immobilized A431 cells as antigen (Wels et al., 1995).

The parental Mab 225 as well as recombinant scFv(225) molecules derived thereof compete the binding of EGF to the EGF receptor thereby inhibiting receptor activation (Fan et al., 1993; Beerli et al., 1994; Wels et al., 1995). In order to test whether the scFv(14E1) domain can also block the binding of EGF, competition experiments were performed. NE1 murine fibroblasts expressing human EGF receptor cDNA (Beerli et al., 1994) were grown in low serum for 16 hr and then treated for 10 min at 37.degree. C. with 10 ng/ml EGF with or without the addition of 20 .mu.g/ml of scFv(14E1)-ETA or scFv(225)-ETA as competitors. Control cells were treated with PBS or scFv-ETA proteins in the absence of EGF. Equal amounts of cell lysates were assayed for their phosphotyrosine content by SDS-PAGE and subsequent immunoblotting with a specific anti-phosphotyrosine monoclonal antibody, followed by incubation with an anti-mouse horseradish peroxidase labeled antibody and chemiluminescent detection. The results are not shown. Treatment of cells with EGF led to a strong increase in the phosphotyrosine content of a protein corresponding in size with the 170 kDa EGF receptor, which was confirmed by reprobing the filter with an anti-EGF receptor serum (data not shown). This EGF-induced activation of the receptor was blocked to a great extent by scFv(225)-ETA whereas scFv(14E1)-ETA completely abolished receptor activation. PBS and scFv-ETA proteins alone had no effect on the phosphotyrosine content of the receptor. The results show that the recombinant scFv(14E1)-ETA similar to scFv(225)-ETA is able to block EGF-induced receptor activation, but is much more potent than the latter at identical concentrations.

In vitro cytotomicic activity and specificity of scFv(14E1)-ETA. Both antibody-toxins, scFv(14E1)-ETA and scFv(225)-ETA inhibit the activation of EGF receptor by EGF. In order to analyze the potential of Mab 225 to interfere with scFv(14E1)-ETA binding, cell killing experiments were carried out. The cytotoxic activity of scFv(14E1)-ETA was tested on A431 cells using an enzymatic assay (Wels et al., 1992 a). The cells were incubated for 40 h with 100 ng/nl (1.5 nM) of scFv(14E1)-ETA or scFv(225)-ETA in the absence or presence of a 100-fold molar excess of Mab 225 and cell viability was measured in comparison to PBS treated cells. The isotype-matched ErbB-2 specific Mab FRP5 (Harwerth et al., 1992) was included as a control. The results are shown in FIG. 3. At the concentration used, both antibody-toxins displayed similar cell killing activity; approximately 90% and 86% of the cells were killed by scFv(14E1)-ETA (FIG. 3, right panel) and scFv(225)-ETA (FIG. 3, left panel), respectively. In the presence of an excess of the specific competitor, Mab 225, the cytotoxic activity was reduced in both cases. No cell killing was observed in the case of scFv(225)-ETA; approximately 33% of the cells were killed in the case of scFv(14E1). An excess of the non-specific Mab FRP5 had no effect on the cytotoxic activity of the scFv-ETA proteins. The results show that the cytotoxic acitivity of scFv(14E1)-ETA is specifically targeted to the EGF receptor, since cell killing can be competed by the EGF receptor specific Mab 225.

In vitro cell killing activity of recombinant toxins specific for the EGF receptor. Since scFv(14E1)-ETA displays a significantly higher affinity for the EGF receptor, cell killing experiments were performed to investigate whether the increased affinity results in enhanced cytotoxicity towards EGF receptor expressing cells. In addition, TGF.alpha.-ETA was included in the experiment, a recombinant fuision toxin which employs the natural EGF receptor ligand TGF.alpha. as a cell targeting domain (Schmidt et al., 1996). The structure of TGF.alpha.-ETA which is schematically shown in FIG. 1A, is very similar to that of the TGF.alpha.-PE40 molecule previously characterized by others (Siegall et al., 1989). The in vitro toxicity of the scFv(14E1)-ETA, the scFv(225)-ETA, and the TGF.alpha.-ETA proteins was tested on three human tumor cell lines. The A431 epidermoid and the MDA-MB468 breast carcinoma cells express between 1 and 2 million EGF receptors per cell, whereas the SKBR3 breast carcinoma cells express approximately 50-fold lower EGF receptor levels on their surface (Wels et al., 1995).

The cells were incubated for 40 h with various concentrations of the EGF receptor specific scFv-ETA and TGF.alpha.-ETA proteins. As a control, the SKBR3 cells which overexpress ErbB-2 were also treated with the ErbB-2 specific antibody-toxin scFv(FRP5)-ETA (Wels et al., 1992 a). The relative number of viable cells in comparison to untreated controls was determined with an enzymatic assay (Wels et al., 1992 a). The results are shown in FIG. 4. A431 cells were very sensitive to the toxins with scFv(14E1)-ETA and TGF.alpha.-ETA showing similar activity (IC.sub.50 below 1 ng/ml) and scFv(225)-ETA being less potent at lower toxin concentrations (IC.sub.50 3 ng/ml) (FIG. 4, upper left panel). Very similar results were obtained with MDA-MB468 cells (FIG. 4, upper right panel). TGF.alpha.-ETA and scFv(14E1)-ETA displayed similar cell killing activity with IC.sub.50 values of approximately 2 and 3 ng/ml, respectively. ScFv(225)-ETA was also cytotoxic for MDA-MB468 cells in a dose dependent fashion, but with a much higher IC.sub.50 value of approximately 40 ng/ml. The EGF receptor specific antibody-toxins scFv(14E1)-ETA and scFv(225)-ETA at concentrations of up to 1 .mu.g/ml did not display significant cell killing activity on SKBR3 cells expressing high amounts of ErbB-2 but only moderate levels of the EGF rececptor (FIG. 4, lower panel). In striking contrast, the recombinant growth factor toxin TGF.alpha.-ETA was cytotoxic for SKBR3 cells at relatively low concentrations with an IC.sub.50 value of approximately 90 ng/ml comparing to an IC.sub.50 of approximately 50 ng/ml for the ErbB-2 specific antibody-toxin scFv(FRP5)-ETA. Similar results were also obtained with T47D human breast tumor cells which like SKBR3 cells express only low levels of the EGF receptor. T47D cells were sensitive to TGF.alpha.-ETA (IC.sub.50 of 105 ng/ml) (Schmidt and Wels, 1996) but highly resistant to scFv(14E1)-ETA and scFv(225)-ETA (IC.sub.50 >1 .mu.g/ml; data not shown). The results show that scFv(14E1)-ETA is highly cytotoxic for tumor cells overexpressing the EGF receptor. Its activity is comparable to that of TGF.alpha.-ETA, but much more potent than that of scFv(225)-ETA. Importantly, both antibody-toxins in contrast to the growth factor toxin are highly selective for EGF receptor overexpressing cells.

In order to analyze the kinetics of scFv-ETA binding to tumor cells and scFv-ETA-mediated cell killing, a time course experiment was carried out. A431 cells were treated for defined time intervals with 100 ng/ml each of scFv(14E1)-ETA and scFv(225)-ETA, respectively. Control cells were treated with the ErbB-2 specific scFv(FRP5)-ETA protein. Unbound scFv-ETA proteins were removed and the cells were incubated in normal growth medium for another 40 h. The relative number of viable cells was determined as described above. The results are shown in FIG. 5. Potent cell killing was achieved after short incubation of the cells with the EGF receptor specific scFv-ETA proteins. While maximal cell killing by scFv(14E1)-ETA was already observed after incubation for 2 min, a similar effect was achieved only after incubation for 30 min with scFv(225)-ETA. In contrast, A431 cells which express low levels of ErbB-2 had to be treated with the ErbB-2 specific scFv(FRP5)-ETA for 100 min to achieve significant cell killing. Our data show that A431 cell killing mediated by EGF receptor specific scFv-toxins follows different kinetics than cell killing by the ErbB-2 specific toxin. This might not only reflect the approximately 50-fold difference in EGF receptor and ErbB-2 expression which could allow a more rapid binding of amounts of anti-EGF receptor toxins sufficient for cell killing, but could also be due to differences in the internalization rates and intracellular pathways of EGF receptor and ErbB-2. The much faster cell killing by scFv(14E1)-ETA in comparsion to scFv(225)-ETA is likely due to the higher affinity of the former. Differences in the internalization and the intracellular routing of toxin-receptor complexes seem unlikely since both scFvs bind to an identical or very similar epitope and do not induce receptor activation upon binding.

Inhibition of TGF.alpha.-ETA binding by scFv proteins. Many human tumor cells which overexpress EGF receptors also synthesize increased amounts of TGF.alpha., which is transported to the cell surface and activates EGF receptors in an autocrine fashion (Derynck et al., 1987; Van de Vijver et al., 1991). This in turn can lead to an increased growth response of the tumor cells. The antagonistic Mab 225 displays growth inhibitory activity on tumor cells in vitro in the absence of active complement or immune effector cells, suggesting that the antitumoral activity of the antibody is mainly due to its ability to block EGF receptor activation and to interrupt autocrine stimulation by TGF.alpha. (Ennis et al., 1989). The very rapid cell killing effect observed after treatment of tumor cells with scFv(14E1)-ETA and scFv(225)-ETA suggests that the inhibition of in vitro tumor cell growth by these proteins is mainly due to their cytotoxic activity. Nevertheless, the ability to interfere with TGF.alpha. binding to the EGF receptor and TGF.alpha.-induced receptor activation might be an additional advantage on autocrine stimulated tumor cells. In order to test the capacitiy of the scFv domains themselves to block TGF.alpha. binding, competition experiments with scFv(14E1) and scFv(225) proteins lacking the cytotoxic ETA domain were performed.

A431 cells were treated for 10 min at 37.degree. C. with 500 ng/ml of TGF.alpha.-ETA with or without the addition of 50 .mu.g/ml of scFv(14E1) or scFv(225) as competitors. Control cells were treated with PBS or 20 ng/ml EGF. Equal amounts of cell lysates were assayed for their phosphotyrosine content by SDS-PAGE and subsequent immunoblotting with a specific anti-phosphotyrosine monoclonal antibody, followed by incubation with an anti-mouse horseradish peroxidase labeled antibody and chemiluminescent detection. The results are not shown in. Treatment of cells with TGF.alpha.-ETA led to a strong increase in the phosphotyrosine content of the EGF receptor comparable to treatment with EGF. No TGF.alpha.-ETA-induced activation of the receptor was observed with scFv(14E1) as a competitor, whereas scFv(225), while also active as a competitor, was unable to completely abolish EGF receptor activation. PBS had no effect on the phosphotyrosine content of the receptor. Similar to the results obtained with scFv-ETA proteins the data show that the recombinant scFv domains of Mab 14E1 and 225 alone are able to block ligand-induced receptor activation, but that scFv(14E1) is much more potent than scFv(225).

To analyze the competition of TGF.alpha. binding by scFv(14E1) and scFv(225) in more detail, an experiment was carried out to block the cytotoxic effect of TGF.alpha.-ETA on several human tumor cell lines by addition of the scFv proteins. MDA-MB468, A431, and SKBR3 cells were incubated for 40 h with 100 ng/ml of TGF.alpha.-ETA with or without the addition of 10 .mu.g/ml of scFv(14E1) or scFv(225), and cell viability was measured in comparison to PBS treated cells as described above. The results are shown in FIG. 6. Both scFv proteins potently inhibited TGF.alpha.-ETA cytotoxicity on SKBR3 cells expressing moderate levels of the EGF receptor. Without competitor, approximately 56% of the cells were killed whereas in the presence of either scFv protein cell killing was reduced to less than 10% (FIG. 6, lower panel). On A431 cells overexpressing the EGF receptor, cell killing was almost complete after incubation with TGF.alpha.-ETA (94% cell killing) (FIG. 6, middle panel). Addition of an excess of scFv(225) hardly influenced the cytotoxic effect of TGF.alpha.-ETA resulting in approximately 90% cell killing. In contrast, scFv(14E1) was very potent in blocking the cytotoxic activity of TGF.alpha.-ETA on A431 cells reducing cell killing to approximately 28%. Similar results were obtained MDA-MB468 cells (FIG. 6, upper panel). TGF.alpha.-ETA treatment alone resulted in approximately 85% cell killing whereas the addition of scFv(225) or scFv(14E1) reduced the cell killing by TGF.alpha.-ETA to approximately 63% and 15%, respectively. The results show that scFv(14E1) is a much more potent competitor of TGF.alpha. binding than scFv(225), and that scFv(14E1) can protect cells from attack by TGF.alpha.-ETA over a prolonged period of time.

Monoclonal Antibody 14E1 Binds to Full-Length and Variant EGFR. DH-EGFR-E and DH-E.DELTA.801-P are NIH3T3 cell lines stably transfected with human full-length and variant EGFR cDNAs, respectively (Hills, D., Rowlinson-Busza, G., and Gullick, W. J. Specific targeting of a mutant, activated EGF receptor found in glioblastoma using a monoclonal antibody. Int. J. Cancer, 63: 537-543, 1995). The phosphotyrosine content of EGFR proteins in DH-EGFR-E and DH-E.DELTA.801-P cells was determined. The cells were grown in a low serum for 16 hr and then treated for 10 min at 37.degree. C. with 20 ng/ml EGF or with PBS. Equal amounts of cell lysates were assayed for their phosphotyrosine content by SDS-PAGE and subsequent immunoblotting with a specific anti-phosphotyrosine monoclonal antibody, followed by incubation with an anti-mouse horseradish pedroxidase labeled antibody and chemiluminescent detection. The results are not shown in. In DH-E.DELTA.810-P cells the variant EGFRvIII displayed high constitutive phosphotyrosine levels which were very similar in the absence or in the presence of EGF. In DH-EGFR-E cells full-length EGFR did not contain measurable amounts of phosphotyrosine in the absence of EGF, but was strongly phosphorylated on tyrosine residues upon EGF treatment.

The murine Mab 14E1 binds to the extracellular domain of human EGFR. The ability of Mab 14E1 to recognize EGFRvIII which lacks amino acids 6 to 273 of the extracellular domain was tested in immunoblot experiments. Equal amounts of cell lysates from DH-EGFR-E and DH-E.DELTA.801-P cells, or from DH-L-2 control cells transfected with an empty vector construct (Hills, D., Rowlinson-Busza, G., and Gullick, W. J. Specific targeting of a mutant, activated EGF receptor found in glioblastoma using a monoclonal antibody. Int. J. Cancer, 63: 537-543, 1995) were separated by SDS-PAGE, electroblotted onto PDVF membranes, and analyzed with Mab 14E1 (data not shown). As a control a similar experiment was performed using Mab DH8.3 which is specific for EGFRvIII (Hills, D., Rowlinson-Busza, G., and Gullick, W. J. Specific targeting of a mutant, activated EGF receptor found in globlastoma using a monoclonal antibody. Int. J. Cancer, 63: 537-543, 1995) (data not shown). As expected, Mab DH8.3 recognized the EGFRvIII protein expressed by DH-E.DELTA.801-P cells but did not bind to proteins from DH-EGFR-E and DH-L-2 cells. In contrast, under non-reducing conditions Mab 14E1detected both, full-length human EGFR and EGFRvIII expressed by DH-EGFR-E and DH-E.DELTA.801-P cells, respectively, but did not bind to control proteins. When cellular lysates were applied to the gel under reducing conditions, Mab 14E1 failed to detect the EGFR proteins (data not shown).

EGFRvIII Expressing Cells are Differentially Sensitive to the Antibody-Toxin scFv(14E1)ETA and the Growth Factor Toxin TGF.alpha.-ETA. Mab 14E1 and recombinant single-chain Fv derivatives of Mab 14E1 compete the binding of the natural EGFR ligands EGF and TGF.alpha. to the receptor, suggesting that the antibody binds to an EGFR epitope very close to the binding site of the growth factors.

The toxicity of scFv(14E1)-ETA and TGF.alpha.-ETA proteins was tested on the NIH3T3 cell lines DH-EGFR-E and DH-E.DELTA.801-P. The cells were incubated for 40 h with various concentrations of the scFv(14E1)-ETA and TGF.alpha.-ETA proteins. As a control, the cells were also treated with the ErbB2 specific antibody-toxin scFv(FRP5)-ETA (Wels, W., Harwerth, I. M., Mueller, M., Groner, B., and Hynes, N. E. Selective inhibition of tumor cell growth by a recombinant single-chain antibody-toxin specific for the erbB-2 receptor. Cancer Res., 52: 6310-7, 1992). The relative number of viable cells in comparison to untreated controls was determined with an enzymatic assay as described in Materials and Methods. The results are shown in FIG. 7. On DH-EGFR-E cells scFv(14E1)-ETA and TGF.alpha.-ETA displayed cell killing activities in a similar range with IC.sub.50 values of 19 ng/ml and 42 ng/ml, respectively (FIG. 7, upper left panel). In contrast, DH-E.DELTA.801-P cells were sensitive only to high concentrations of TGF.alpha.-ETA (IC.sub.50 of 0.7 .mu.g/ml) but were extremely sensitive to low concentrations of scFv(14E1)-ETA (IC.sub.50 of 0.2 ng/ml) (FIG. 2, upper right panel). As expected, none of the cell lines were sensitive to the ErbB2 specific antibody-toxin. These data show that DH-E.DELTA.801-P cells which express EGFRvIII are approximately 3,500 times more sensitive to the scFv(14E1)-ETA antibody-toxin than to the TGF.alpha.-ETA growth factor toxin, and are approximately 100 times more sensitive to scFv(14E1)-ETA than DH-EGFR-E cells expressing full-length human EGFR.

A similar cell killing experiment was carried out using DH-L-2 control cells which do not express human EGFR. Neither the EGFR specific toxins TGF.alpha.-ETA and scFv(14E1)-ETA nor the ErbB2 specific control protein scFv(FRP5)-ETA showed detectable cytotoxic activity on DH-L-2 cells (FIG. 2, lower panel). These data show that the activity of scFv(14E1)-ETA and TGF.alpha.-ETA proteins on DH-E.DELTA.801-P and DH-EGFR-E NIH3T3 transfectants is dependent on the presence of the human EGFR proteins.

Increased Sensitivity to scFv(14E1)-ETA is Dependent on EGFRvIII Expression. As shown in Example 6, DH-EGFR-E and DH-E.DELTA.801-P cells express different levels of full-length and variant EGFR. In order to test whether the differential sensitivity of these cells to scFv(14E1)-ETA can be explained by the amount of EGFR proteins present, DH-E.DELTA.801-5 (Hills, D., Rowlinson-Busza, G., and Gullick, W. J. Specific targeting of a mutant, activated EGF receptor found in glioblastoma using a monoclonal antibody. Int. J. Cancer, 63: 537-543, 1995), another NIH3T3 cell line expressing only low levels of EGFRvIII was analyzed. First, the relative expression levels of EGFR proteins in DH-EGFR-E, DH-E.DELTA.801-P, and DH-E.DELTA.801-5 cells were determined by quantitative immunoblotting using the rabbit polyclonal antibody 12E which recognizes both, full-length and variant EGFR (Hills, D., Rowlinson-Busza, G., and Gullick, W. J. Specific targeting of a mutant, activated EGF receptor found in glioblastoma using a monoclonal antibody. Int. J. Cancer, 63: 537-543, 1995., Gullick, W. J., Downward, J., and Waterfield, M. D. Antibodies to the autophosphorylation sites of the epidermal growth factor receptor protein-tyrosine kinase as probes of structure and function. EMBO J., 4: 2869-77, 1985). The results are not shown. Equal amounts of cell lysates from DH-EGFR-E cells, DH-L-2 control cells, and DH-E.DELTA.801-P and DH-E.DELTA.801-5 cells were analyzed by SDS-PAGE and inmuunoblotting with 12E antibody. As determined by densitometry, DH-E.DELTA.801-P and DH-EGFR-E cells express approximately 11 times and 3.5 times higher EGFR levels than DH-E.DELTA.801-5 cells, respectively.

The cytotoxic activity of scFv(14E1)-ETA and TGF.alpha.-ETA proteins on DH-E.DELTA.801-5 cells was analyzed. The cells were incubated for 40 h with various concentrations of the recombinant toxins and the relative number of viable cells was determined as described above. The results are shown in FIG. 10. DH-E.DELTA.801-5 cells were highly sensitive to scFv(14E1)-ETA (IC.sub.50 of 2 ng/ml), but resistant to low concentrations of TGF.alpha.-ETA (IC.sub.50 >1 .mu.g/ml). In comparison to the DH-E.DELTA.801-P cells described above, DH-E.DELTA.801-5 cells were 10 times less sensitive to scFv(14E1)-ETA which is in good agreement with the approximately 11 times lower EGFRvIII expression. In contrast, in comparison to DH-EGFR-E cells, DH-E.DELTA.801-5 cells displayed 10-fold higher sensitivity to scFv(14E1)-ETA despite the 3.5 times lower level of EGFR proteins.

These results show that the sensitivity of EGFRvIII expressing cells towards scFv(14E1)-ETA is dependent on the amount of variant receptor present. However, the increased sensitivity of DH-E.DELTA.801-P and DH-E.DELTA.801-5 cells to scFv(14E1)-ETA in comparison to DH-EGFR-E cells cannot be explained by the total amount of EGFR proteins in these cells but is dependent on the presence of EGFRvIII.

The present invention relates to a single chain antibody-toxin, scFv(14E1)-ETA, specific for the EGF receptor and/or its variants which is based on the independently isolated Mab 14E1. This single chain antibody-toxin is well suited for targeting cells via a surface molecule such as the EGF receptor: The target-cell recognition domain binds with high affinity to its cognate receptor. Furthermore, it shows a high selectivity for EGF receptor overexpressing tumor cells and therewith minfimize unwanted side effects on normal tissues that could express low levels of the target antigen, and the cytotoxic domain of the anti-tumor toxin is catalytically active so that the molecule is effective at low concentrations.

The antagonistic activity of scFv molecules could be employed directly to inhibit tumor cell growth in the absence of a cytotoxic effector domain in a way very similar to the parental antibodies. ScFv proteins of approximately 27 kDa due to their much smaller size penetrate into solid tumors much faster than intact Mabs (approximately 150 kDa). This is an advantageous especially for tumor imaging with radiolabeled compounds, where the rapid clearance of scFv molecules from the circulation of usually only several minutes is desirable.

scFv(14E1)-ETA is almost as potent as TGF.alpha.-ETA in tumor cell killing, since the antibody-toxin cannot activate the EGF receptor and is therefore limited to the passive internalization of the target antigen in order to reach the cytosol, the site of toxin activity. Since in contrast to TGF.alpha.-containing toxins scFv(14E1)-ETA lacks significant killing activity on cells expressing low EGF receptor levels, it will be more tolerable upon systemic administration and allow a broader range of therapeutic applications. ScFv(14E1)-ETA is a promising and very potent novel antitumoral reagent.