Potentiation of Cytotoxic Conjugates

1. Field of the Invention

The present invention relates generally to cancer therapy and, more particularly, to the potentiation or enhancement of cytotoxic conjugates of monoclonal antibodies (MoAbs) and ribosomal inhibiting proteins (RIPs) by the co-administration of a second, unconjugated monoclonal antibody.

Colorectal cancer is the second most common cause of death from malignancy in the Western world. The American Cancer Society estimates that there were 138,000 new cases of colorectal cancer and 59,900 patients died from the disease in 1985. The outlook for patients with colorectal disease has remained essentially unaltered over the last 30 years, with the five-year survival being on the order of 30%. A major factor contributing to this is the lack of effective treatment for the disease once it has spread beyond the bowel wall, since surgery during the early stages of the disease offers the only prospect of cure.

Unfortunately, the majority of patients have disseminated disease at the time of initial surgery, primarily with overt or occult hepatic and lymph node metastases.

Despite numerous trials of chemotherapy, the only single agent shown to have any significant effect is 5-fluorouracil; its response rate is inadequate and treatment with it rarely influences the ultimate outcome. Combination chemotherapy and intrahepatic arterial infusion are being investigated but currently do not represent effective treatments.

Cancer of the ovary accounts for roughly 5% of all cancers in women and is the sixth leading cancer in women. Although surgery is curative if this lesion is detected early enough, the mortality associated with this disease has not improved appreciably in the last 25 years.

Ultrasound, laparoscopy or peritoneoscopy, and CAT scan are of limited value in the diagnosis of ovarian carcinoma. Serum markers such as carcinoembryonic antigen and placental alkaline phosphatase as well as some newly defined antigens are found in the blood of some patients with adenocarcinoma, although there is no universal marker. Surgery with biopsy is the only definitive way of diagnosing ovarian carcinoma.

Surgery is currently the only cure for ovarian carcinoma, and is only curative if the tumor has not spread. Radioisotope implants, x-ray irradiation, and chemotherapy are of limited use in the management of ovarian carcinoma.

The most important prognostic indication is the extent of spread of the tumor at the time of diagnosis and surgery. Stage I ovarian carcinoma (growth limited to the ovaries) has an overall 5-year survival rate of approximately 80% Stage II (growth involving ovaries with pelvic extension) has a 5-year survival rate of 40%, Stage III (growth involving ovaries with extension to small bowel or omentum) has a 5-year survival rate of 10%, and Stage IV (distant metastases) has a 5-year survival rate of less than 5%. The relative 5-year survival rate for ovarian carcinoma is 37% (1973-80), relatively the same as for 1960-63 (32%).

Imaging studies of ovarian carcinoma with radiolabelled MoAbs have been performed to a limited degree in human and animal systems. Tumors as small as 1 mm in diameter in mice bearing xenografts of human ovarian cancer have been imaged. The same study also detected tumors in 8 of 10 patients with ovarian cancer.

Osteogenic sarcoma (OS) is the most common primary bone tumor. Although surgery can be curative if this lesion is detected early enough, the usual course of this disease in 80%-85% of patients is multiple pulmonary metastases and death within two years of diagnosis. These metastases are often present but usually not large enough to be seen at diagnosis of the primary tumor.

Amputation is the treatment of choice for OS. Limb salvage procedures have been performed, such as en-block resection and prosthetic replacement. Overall survival for limb salvage is either as poor as worse than with amputation.

Radiation treatment has not been shown to prevent OS metastases.

Most recent trials note survival rates of over 50% at 5 years for patients treated by widely diverse adjunctive methods. The question is whether all of these widely diverse forms of treatment are effective or if there is a change in the natural history of the disease.

Imaging studies of OS with radiolabelled MoAbs have been performed to a limited degree in human and animal systems. Human OS xenografts have been imaged in nude mice using an anti-OS MoAb labelled with .sup.131 I.

Because of the extent of these three forms of cancer, there is a need for new compositions and methods to treat primary, recurrent and metastatic disease.

2. Description of the Relevant Literature

Embleton et al., Br. J. Cancer (1981) 43:582-7, report preparation of a hybridoma against osteogenic sarcoma cell line 791T, and binding properties of the resulting MoAb with respect to a variety of cell types.

Farrands et al., Lancet (1982) 2:397-400, describe radioimaging of human colorectal cancers using MoAb 791T radiolabelled with .sup.131 I.

Pimm et al., Int. J. Cancer (1982) 30:75-85, describe in vitro localization of osteogenic sarcoma xenografts in mice using radiolabelled 791T MoAbs.

Price et al., Br. J. Cancer (1982) 46:610-10, report on complement-dependent cytotoxicity of two anti-791T osteogenic sarcoma MoAbs against human tumor cell lines, and failure of one of these MoAbs to inhibit 791T tumor xenografts in mice.

Embleton et al., Br. J. Cancer (1983) 47:43-9, describe coupling of the drug vindesine to 791T MoAbs, and in vitro cytotoxicity testing of the conjugate.

Garnette, et al., Int. J. Cancer (1983) 31:661-70, describe in vitro cytotoxicity testing of a 791T MoAb methotrexate conjugate.

Baldwin, et al., Bull.Cancer (1983) 70:132-6, describe the use of radiolabelled 791T MoAb to detect colorectal carcinoma, and evaluate the MoAb for targeting anti-tumor agents including cytotoxic drugs and immunomodulating agents.

Pelham, et al., Cancer Immunol. Immunother. (1983) 15:210-6, describe coupling of interferon to the MoAb 791T, and the retention of biologic activity if both entities of the conjugate.

Baldwin and Pimm. Cancer Metastasis Rev. (1983) 2:89-106, summarize a series of radioimaging experiments using .sup.131 I labelled 791T MoAbs to localize human tumor xenografts in mice.

Farrands et al., J. Bone Joint Surg. (1983) 65:638-4, report on detection of a primary osteogenic sarcoma in a patient using .sup.131 I labelled 791T MoAbs.

Price et al., Scand. J. Immunol. (1983) 18:411-20, report on the identification of an antigen on mitogen-stimulated peripheral blood mononuclear cells using the MoAb 791T, and the characterization of this antigen as being identical to the 791T MoAb-defined antigen found on osteogenic sarcoma cells.

Embleton et al., Br. J. Cancer (1984) 49:559-65, describe treatment of osteogenic sarcoma cells with a methotrexate/791T MoAb conjugate at toxic concentrations which allowed "escape" of some tumor colonies, which were then evaluated for growth potential.

Armitage et al., Br. J. Surg (1984) 71:407-12, report on the imaging of gastrointestinal cancers in patients and discuss radiolabelled 791T MoAb uptake in malignant and non-malignant tissue.

Pimm and Baldwin, Eur J. Cancer Clin. Oncol. (9184) 20:515-24, describe the extent and rate of localization of radiolabelled 791T MoAb in osteogenic sarcoma xenografts in mice.

Price et al., FEBS Lett. (1984) 171:31-5, report on the characterization of the cell surface antigen p72 with which the MoAb 791T reacts.

Price et al., Br. J. Cancer (1984) 49:809-12, review radioimaging studies using 791T and discuss localization of the antibody within the tumor.

Gallego et al., Int. J. Cancer (1984) 33:737-44, report on the coupling of the drug daunomycin to the MoAb 791T, using four different coupling procedures, and evaluate the four conjugates for cytotoxicity to tumor cells.

Flannery et al., Eur. J. Cancer Clin. Oncol. (1984) 20:791-8, describe coupling of interferon to the MoAb 791T, and the activation of NK cells by this conjugate.

Campbell et al., Int. J. Cancer (1984) 34:31-7, describe the analysis of expression on various cell lines of the antigen with which the MoAb 791T reacts.

Baldwin et al., Symp. Fundam. Cancer Res. (1983) 36:437-55, describe radioimaging of primary and metastatic colorectal carcinomas in patients using radiolabelled 791T MoAb.

Embleton et al., Behring Inst. Mitt. (1984) 74:108-11, summarize cytotoxicity testing of a vindesine/791T MoAb conjugate and a methotrexate/791T MoAb conjugate.

Embleton et al., Behring Inst. Mitt. (1984) 74:35-8, summarize reactivity of two anti-human tumor MoAbs, 791T and C14/1/46.

Pimm et al., Behring Inst. Mitt. (1984) 74:80-6, describe detection of primary and metastatic colorectal carcinoma in patients using .sup.131 I labelled 791T MoAb.

Williams et al., Clin. Oncol. (1984) 10:375-81, describe detection of primary and metastatic mammary carcinoma in patients using .sup.131 I labelled 791T MoAb.

Symonds et al., Br. J. Obstet Gynaecol. (1985) 92:270-6, describe preliminary results of detection of ovarian tumors in patients using .sup.131 I labelled 791T MoAb.

Rowland et al., Cancer Immunol. Immunother. (1985) 19:1, report formation of four vindesine-MoAb conjugates, including the MoAb 791T, and compare the conjugates, free vindesine, and free MoAb for their ability to inhibit tumor cell growth.

Pimm et al., Cancer Immunol. Immunothen. (1985) 19:18-21, describe localization of a primary osteogenic sarcoma tumor in situ using .sup.131 I-labelled 791T MoAb, and subsequent localization of xenografts of the tumor in mice.

Perkins et al., Eur. J. Nucl. Med. (1985) 10:296-301, report on a method of labelling the MoAb 791T with .sup.111 In.

Pimm et al., J. Nucl. Med. (1985) 26:1011-23, describe characteristics of the blood survival of .sup.131 I and .sup.111 In-labelled 791T MoAbs in patients with colorectal carcinoma, ovarian carcinoma, and osteogenic sarcoma in tumor imaging studies.

Pimm et al, Eur. J. Nucl. Med. (1985) 11:300-4, describe comparison of the blood, tumor and wholebody levels of .sup.131 I and .sup.111 In labelled 791T MoAb in mice with human tumor xenografts.

Pimm et al., ICRS Med. Sci. (1985) 13:499-500, describe the production of anti-idiotypic antibodies to 791T MoAb following administration of radiolabelled 791T MoAb in a series of patients.

Pimm et al., ICRS Med. Sci. (1985) 13:366-7, describe a solid-phase micro-radioimmunoassay as a means of measuring MoAb reactivity with antigens, using 791T as an example.

Armitage et al., Nucl. Med. Commun. (1985) 6:623-31, describe imaging of primary and metastatic colorectal carcinoma using .sup.111 In labelled 791T MoAb.

Durrant et al., Br. J. Cancer (1986) 53:37-45, report on the localization of the MoAb 791T within xenografts derived from colorectal adenocarcinomas.

Garnett and Baldwin, Cancer Res. (1986) 46:2407-12 describe improved synthesis of a drugcarrier-MoAb conjugate using methotrexate and 791T.

Ramakrishnan and Houston, Science (1984) 223:58-61, describe the potentiation of immunotoxins directed against human acute lymphoblastic leukemia by chloroquine.

Akiyama et al., Cancer Res. (1985) 45:1005-07, report the potentiation of the cytotoxic activity against human tumor cells of toxic conjugates of Pseudomonas exotoxin with anti-transferrin receptor antibody or epidermal growth factor up to 10 to 20 fold by the calcium antagonists verapamil, D-600, and diltiazem and by the lysosomotropic agent .beta.-glycylphenylnaphthylamide.

Uckun et al., Blut (1985) 50:19-23 describe the potentiation by mafosfamid (ASTA Z 7557) of the ex vivo efficacy of a T-cell directed immunotoxin containing pokeweed antiviral protein (PAP).

Novel compositions and methods are provided for the treatment of cancer employing monoclonal antibodies (MoAbs) conjugated to a toxin. According to the present invention, MoAbs defining epitopes on either a tumor associated glycoprotein antigen of about 72 kD m.w. or on carcinoembryonic antigen conjugated to a ribosomal inhibiting protein, or the like, including the A chains of the toxic lectins ricin and abrin, are employed either alone or in combination with other conjugates, compositions, treatments, or the like, as cytotoxic agents in the treatment of various cancers including, but not limited to, colorectal carcinoma, ovarian carcinoma and osteogenic sarcoma. For some of these compositions, an enhanced or potentiated efficacy is observed when the conjugates are administered along with an unconjugated monoclonal antibody.

FIG. 1 is a graphic illustration of the results obtained in the therapy of mice bearing xenografts of osteogenic sarcoma 791T cells with XMMCO-791-RTA;

FIG. 2 is a graphic illustration of the results obtained in the therapy of mice bearing xenografts of colorectal carcinoma C170 cells with XMMCO-791-RTA at a dose of 5 mg/kg/day;

FIG. 3 is a graphic illustration of the results obtained in the therapy of mice bearing xenografts of colorectal carcinoma C170 cells with XMMCO-791-RTA at a dose of 3 mg/kg/day; and

FIG. 4 is a graphic illustration of the influence of free RTA and XMMCO-228-RTA on the growth of colorectal carcinoma LS174T cell xenografts .

The present invention employs cytotoxic conjugates for the treatment of various forms of cancer, including colorectal carcinoma, ovarian carcinoma and osteogenic sarcoma. These novel conjugates are comprised of monoclonal antibodies or binding fragments thereof, collectively termed immunoglobulins, bound to a cytotoxin. These compositions are administered to a cancer cell host in order to destroy cancer cells while doing minimal damage to normal tissue.

The immunoglobulins of the present invention are employed as targeting agents for directing cytotoxic agents to specific cancer cells within a cancer cell host. According to the present invention, immunoglobulins which define an epitope on a 72 kilodalton (kD) glycoprotein antigen and immunoglobulins which define an epitope on carcinoembryonic antigen (CEA) are employed as targeting agents. One or both of these antigens are present on a variety of cancers including, but not limited to, colorectal carcinoma, ovarian carcinoma and osteogenic sarcoma.

These targeting immunoglobulins are bound to a variety of cytotoxic agents, generally ribosomal inhibiting proteins, more particularly, the A chain of ricin or the A chain of abrin. The A chain of ricin has been successfully employed in cytotoxic conjugates as disclosed in U.S. Pat. No. 4,590,071, the disclosures of which are hereby incorporated by reference.

The immunoglobulin and toxin are generally bound by a covalent bond, more particularly a disulfide bond, but may be joined by any chemical bond which allows the toxin to travel to the target cell with the immunoglobulin. Other methods for achieving such a bond are well-known to those skilled in the art. The only criteria is that the bond must be achieved in a manner which does not significantly decrease the binding affinity of the immunoglobulin for its epitope.

The present cytotoxic conjugates may be administered to a cancer cell host either singly or in a cocktail containing two or more conjugate formulations, other chemotherapeutic agents, compositions, or the like. Cocktails are particularly important in the treatment of heterogeneous tumor cell populations wherein targeting of multiple antigens is critical.

The cytotoxic conjugates of the present invention may be administered to a cancer cell host by any convenient method. Pharmaceutical compositions employing the subject conjugates may be administered parenterally, i.e., intravenously, intraparitoneally, or the like. Thus, this invention provides compositions for parenteral administration which comprise a solution of pyrogen free, cytotoxic conjugates, or a cocktail thereof, dissolved in an acceptable carrier, preferably an aqueous carrier. A variety of aqueous carriers can be used, e.g., water, buffered water, 0.4% saline, 0.3% glycine, or the like. These solutions are sterile and generally free of particulate matter. These compositions may be sterilized by conventional, well-known filtration sterilization techniques. The compositions may contain pharmaceutically acceptable auxiliary substances as required to approximate physiological conditions such as pH adjusting and buffering agents, toxicity adjusting agents, or the like, for example, sodium acetate, sodium chloride, potassium chloride, potassium chloride, calcium chloride, sodium lactate, etc.

The compositions of the present invention can be lyophilized for storage and reconstituted in a suitable carrier prior to use. This technique has been shown to be effective with conventional immune globulins and art-known lyophilization and reconstitution cell kill, is increased for a given dose of immunotoxin. The unconjugated immunoglobulin will generally be related to the conjugate immunoglobulin in that it will bind a different epitope of the same cell surface antigen. The immunoglobulins need not be related in that any unconjugated immunoglobulin, or other ligand, that binds target cell surface receptors which trigger endocytosis of the toxic moeity may be employed with equal effect. It is believed that the co-administration of unconjugated immunoglobulin potentiates the immunotoxin by enhancing endocytosis of the toxic moeity. This is achieved by binding multiple cell surface receptors responsible for the triggering of endocytosis.

The dose of unconjugated immunoglobulin will vary widely depending upon the immunotoxin employed, choice of unconjugated immunoglobulin, nature and extent of the tumor, and the like. The dose will generally be in the range of 0.001 mg/kg/day to 100.00 mg/kg/day, usually about 0.01 mg/kg/day to 10.0 mg/kg/day and, more particularly, about 0.1 to 1 mg/kg/day. The ratio of unconjugated immunoglobulin to conjugate will also vary widely, generally from about 10,000-0.001 to 1, usually about 1000-0.01 to 1 and, more particularly, about 100-0.1 to 1.

The following examples are offered by way of illustration and not limitation.

A. Production of XMMCO-791 Hybidomas

Balb/c mice (Bantin and Kingman, U.K.) were immunized with cultured 791T cells according to the following schedule:

1. Day 0: 1.times.10.sup.7 cells were given intraperitoneally (IP) in Hank's balanced salt solution (HBSS);

2. Day 7: 1.times.10.sup.7 cells were given IP in HBSS; and

3. Day 21: 2.times.10.sup.6 cells were given by intracardiac injection.

The 791T cell line used for immunization was obtained from Dr. Vera S. Byers, University of California, San Francisco. 791T is a human osteogenic sarcoma cell line isolated from the pulmonary metastasis of a 19-year old Caucasian male. The cell line was characterized as malignant by its ability to grow as a tumor xenograft in nude mice, its lack of contact inhibition in culture and its abnormal nuclear and cytoplasmic morphology.

On day 26 after the initial immunization, spleen cells from immunized mice were aseptically removed. Following procedures as outlined elsewhere (Galfre et al., Nature (1977) 226:550, which is incorporated by reference), 5.times.10.sup.7 spleen cells were fused with an equal number of P3/NSI/1-Ag4-1 cells (A.T.C.C. Accession No. TIB-18) using polyethylene glycol 4000. The P3/NSI/1-Ag4-1 cell line is a non-secreting mouse myeloma cell line of Balb/c origin. This line is resistant to 8-azaguanine and lacks the enzyme hypoxanthine quanine phosphoribosyl transferase so that hypoxanthineaminopterin-thymidine (HAT) medium beocmes toxic, providing a convenient method of selecting unfused parental myeloma cells from the hybridoma population.

Cloning was performed by plating cells at low density in 0.3% agar in RPMI 1640 medium supplemented with 50 mg/ml gentomycin and hypoxanthine-aminopterinthymidine (HAT) with 10% fetal calf serum, and overlaying a gel of 0.5% agar in the same medium. Spherical colonies were removed by Pasteur pipette. Recloning was performed by the limiting dilution technique, in which less than one hybridoma per three wells of a 96-well microtiter plate was plated in the presence of 2.times.10.sup.4 peritoneal exudate cells from unimmunized WAB/Not rats. Medium for regular maintenance was identical to the above, except that no aminopterin was included (HT medium).

The antibodies secreted by each clone were assayed for the appropriate specificity by radioimmune binding assay (Embleton et al., Br. J. Cancer (1981) 43:582-87). One clone, designated XMMCO-791 was found to stably secrete monoclonal antibody; the antibody was determined to be of immunoglobulin class IgG2b. Hybridoma XMMCO-791 is presently on deposit with the American Type Culture Collection (A.T.C.C.), 12301 Parklawn Dr., Rockville, Md. 20852, U.S.A. The deposit was made on August 14, 1986, and given A.T.C.C. Accession No. HB 9173.

Specific pathogen free (SPF) Balb/c mice (Charles River) were used to culture the hybridoma intraperitoneally. Approximately 2.times.10.sup.6 XMMCO-791 hybridoma cells in 0.5 ml of phosphate buffered saline (PBS) pH 7.2 were injected I.P. into pristane primed mice 7 to 21 days post priming. Priming was performed by injecting SPF Balb/c mice I.P. with 0.5 ml Pristane (2,6,10,14-tetramethylpentadecane). The resultant ascites fluid, collected 10-14 days post injection of hybridomas, contained 1.2 to 1.6 mg/ml of the antibody, as determined by double immunodiffusion.

The antibody in ascites fluid was purified following sterile filtration through a 0.22.mu. filtering system by loading onto a Staphylococcal protein A column (Staph protein A) and allowed to bind at 4.degree. C., pH 7.5. The effluent was monitored spectrophotometrically until there was no further change of absorbance. The column was then subjected to pH gradient elutions (pH 5.0 and pH 3.0, 0.1 M sodium citrate buffer) and the monoclonal antibody collected. Purified antibody was stored frozen following aliquoting and quick freezing.

B. Antibody Binding to Cell Lines

The binding of XMMCO-791 to a large number of cultured human tumor cells and cultured human fibroblasts was compared by various assays to define tumor specific binding. These assays are detailed below. Studies show that there is little binding to normal fibroblasts, granulocytes, red blood cells, or lymphoblastoid cell lines; but there is significant binding to osteogenic sarcoma, colorectal and ovarian carcinoma. Studies described later demonstrate significant binding to freshly disaggregated human colorectal and ovarian carcinoma tumors.

1. .sup.125 I Protein A-Binding Assay

Target cells in round-bottomed microtest plates were incubated on ice with hybridoma supernatant. The cells were washed and incubated for an additional 1 hour on ice with Protein A (Pharmacia) labeled with .sup.125 I. The cells were washed and dried down. After spraying with a plastic film (Nobecutane), the wells were separated with a band saw and the bound .sup.125 I measured in a gamma counter.

XMMCO-791 hybridoma supernatants were reacted with cultured cells derived from a range of human tumors and normal fibroblasts. Representative data are summarized in Table I. In tests with 36 malignant cell lines and 9 fibroblast cell lines, XMMCO-791 antibody reacted with 7/13 osteogenic sarcomas; 1/3 lung carcinomas, 1/1 prostate carcinomas, and 1/1 cervical carcinomas. 3/5 colorectal carcinomas also reacted strongly with binding ratios (BR) up to 33.40. Of the fibroblast lines, 1/9 reacted weakly (BR 2.14). Tests were carried out with fibroblasts derived from the donor of the tumor cell line 791T and these target cells did not react with XMMCO-791 (BR 0.99). Peripheral blood mononuclear cells were also negative.

2. Direct Binding Assay with .sup.125 I-Labeled XMMCO791 Antibody

Tissue culture-derived target cells were aliquoted into round-bottomed microtest plates, sedimented by centrifugation, and incubated for 60 minutes with aliquots of .sup.125 I-XMMCO-791. The cells were washed and dried down. The wells were separated from the plate and bound radioactivity counted (Price MR, et al., 1982; Pimm MV, et al., 1982).

Direct binding of .sup.125 I-labeled XMMCO-791 to a range of tumor cells is shown in Table II and this data is summarized in Table III which expresses target cell binding of XMMCO-791 antibody relative to that with 791T cells. These tests demonstrate that .sup.125 I-XMMCO-791 binds to osteogenic sarcomas 791T and 788T, and to HeLa cells. Binding of .sup.125 I-XMMCO-791 to other cell line derived tumor cells, including lung carcinoma (A427 and A549) and bladder carcinoma (T24), was approximately 10% of that with 791T cells.

3. Flow Cytometry

a. Direct Binding of FITC-Labeled XMMCO-791

Target cells were incubated with saturating doses of fluorescein isothiocyanate-labeled XMMCO-791 (FITC-XMMCO-791) and analyzed by flow cytometry. The binding of FITC-XMMCO-791 is expressed as fluorescence units/cell (Price MR, et al., 1983b; Gallego J. et al., 1984; Roe R. et al., 1985).

Specificity is demonstrated by mixing unlabeled MoAb with FITC-labeled MoAb and demonstrating that the fluorescence per cell decreased. Titrating FITC-XMMCO791 antibody against a fixed number of 791T cells indicated that maximum uptake of labeled antibody was obtained at approximately 0.1 .mu.g antibody per 2.times.10.sup.5 cells.

Table IV summarizes tests in which target cells derived from human tumor cell lines were reacted with FITC-XMMCO-791 and analyzed by flow cytometry. The antibody consistently reacted with osteogenic sarcoma 791T and 788T cells; these cells were used as standards for reference purposes. Ovarian carcinoma PA1 cells showed essentially no reactivity with XMMCO-791. All of the tumors listed bound significant amounts of FITC-XMMCO-791, although this was less than that observed with 791T cells. The C146, C168, and C170 cell lines were recently derived from surgical specimens obtained at the Department of Surgery, University Hospital, Queens Medical Centre, Nottingham, U.K. (Director, Professor J.D. Hardcastle). Since these tests were carried out with FITC-XMMCO-791 under saturating conditions, the number of antibody molecules bound/cell can be determined (Roe R., et al., 1985). Thus, colon tumor derived cells bound between 1.6.times.10.sup.5 and 2.8.times.10Z.sup.5 antibody molecules/cells compared with a value of 6.times.10.sup.5 for osteogenic sarcoma 791T cells.

b. Indirect Membrane Immunofluorescence Assay

Tumor cell lines were incubated with saturating amounts of XMMCO-791 monoclonal antibody, then washed and incubated with FITC-labeled rabbit anti-mouse Ig. The samples were analyzed by flow cytometry (Roe R., et al., 1985). Table V summarizes these tests with results expressed as fluorescence units/cell. The results showed the same pattern as with direct immunofluorescence: maximal binding with osteogenic sarcoma cell lines, intermediate with colon carcinoma cell lines, and low levels with ovarian, bladder, and breast carcinoma cell lines.

The binding of XMMCO-791 monoclonal antibody to tumor cells was also compared with that obtained with a range of other murine monoclonal antibodies. These include antibodies recognizing CEA epitopes (B14B8, 194.2 and 161) and other anti-colon carcinoma antibodies. These test, summarized in Table VI, show that XMMCO-791 binds to 791T cells, whereas none of the other antibodies were reactive with this cell line. Both XMMCO-791 and the anti-CEA antibodies reacted with colon tumor cells (LS174T, C170, HT29) and weakly with bladder carcinoma T24.

4. Complement Dependent Cytotoxicity

The complement dependent cytotoxicity of monoclonal antibody MMCO-791 for target cells derived from cultured tumor cell lines was assayed using a 2 hour, .sup.51 chromium (.sup.51 Cr) release assay (Price MR, et al., 1982).

Tumor cells (5.times.10.sup.6) were labeled with .sup.51 Cr. The cells were aliquoted into microtest plates that contained either medium or diluted monoclonal antibody preparations. As a source of complement, normal rabbit serum was added to each well. Following incubation at room temperature for 2 hours, the supernatants were collected and the percentage release of .sup.51 Cr was calculated for each test sample.

Table VII summarizes tests on the complement mediated cytotoxicity of XMMCO-791 monoclonal antibody preparations. XMMCO-791 antibody was cytotoxic for osteogenic sarcoma derived cell lines 791T, 788T, 20S, and 278T, prostate carcinoma derived cell line EB33, and for HeLa cells, but not for ovarian carcinoma derived cell line PA1 or lung carcinoma A549 cells.

C. Antibody Binding to Human Malignant Cells and Tissues

1. Radioimmunoassay of XMMCO-791 Binding to Tumor Membranes

Monoclonal antibody XMMCO-791 binding to membrane preparations derived from surgically removed tumors was determined using a solid phase radioimmunoassay.

Tumor membrane preparations (ENM) were made by homogenizing fresh surgically removed tumor. Nuclei and cell debris were removed by centrifugation (600 g) and the ENM material sedimented at 105,000 g. ENM preparations attached to polyvinylchloride microtiter plates were used to assay antibody binding, this being detected by uptake of .sup.125 I-labeled rabbit anti-mouse IgG F(ab').sub.2 (Brown A., et al., 1983) (Table VIII). Positive tumor:nontumor ratios were obtained in 3/5 tumors.

2. Flow Cytometry

Binding of monoclonal antibody XMMCO-791 to malignant cells derived from primary and metastatic human tumors has been determined by flow cytometry. For this purpose, tumor cells were brought into suspension by collagenase treatment, and reacted with either XMMCO-791 or with control murine monoclonal antibodies. Cell bound immunoglobulin was then detected with FITC-conjugated rabbit anti-mouse Ig and analyzed by flow cytometry. The analysis was restricted to the size range of malignant cells by appropriate forward angle scatter gating (Durrant LG, et al., 1986a).

XMMCO-791 antibody reaction with tumor cells derived from a primary colon carcinoma and from liver and lymph node metastases was tested. Significant binding of XMMCO-791 with tumor cells from both the primary and metastatic tumor was shown. For comparison, binding tests were carried out at the same time with two other murine monoclonal antibodies, C194 and C14.vertline.1.vertline.46; these MoAbs recognize CEA and Y hapten (Brown A., et al., 1983), respectively. Overall, the reaction with XMMCO-791 was comparable with that obtained with these antibodies.

The results of flow cytometry studies on binding of XMMCO-791 MoAbs to freshly disaggregated ovarian tumors are shown in Table VIIIa. The XMMCO-791 MoAb bound to the cell surface of all but one ovarian tumor. Intensity of staining was weak (MLF<100) in one tumor, moderate (MLFs 100-1000) in 8 and strong (MLF>1000) in one.

3. Characterization of XMMCO-791 Binding Antigen

in Human Primary Colon Carcinoma Tumors

As will be described later, studies carried out at University of Nottingham, U.K., with appropriate regulatory approval, demonstrated that the MoAb XMMCO-791, when injected in vivo, localizes in primary and metastatic human colon cancers. Thus, studies were carried out to identify the antigen to which the antibody is bound in vivo.

Briefly, 200 .mu.g .sup.131 I-XMMCO-791 containing 2 mCi .sup.131 I was injected into patients with primary colon carcinomas, and 1-6 days later normal colon tissue and colon tumors were resected. Tissues were homogenized and the fibrous material was extracted. The extract was then mixed with Affi Gel 10 (BioRad Laboratories, Richmond, Calif.), to which affinity purified goat antimouse IgG antibody has been linked. The pellets contained the bound .sup.131 I-labeled material from the tissue extract. This was reacted with .sup.125 I to label tissue protein.

After washing, the Affi Gel pellet was treated with sodium isothiocyanate to release bound .sup.131 I-labeled XMMCO-791 antibody, either free or complexed with antigen.

After purification by gel chromatography, the extract was incubated with Sepharose-Protein A to isolate Ig containing material. These products were removed from the gel and SDS-Polyacrylamide-Gel-Electrophoresis (SDS-PAGE) was carried out. The results are summarized in Table IX.

The apparent molecular weights of two of the bands at 60 kD and 25 kD corresponded to those for heavy and light chains of the XMMCO-791 IgG2b antibody. The third band at 72 kD corresponded to that of the XMMCO-791-defined antigen isolated from 791T cells. The conclusion from the studies is that following infusion of XMMCO-791 antibody into colorectal carcinoma patients, the antibody localizes within the tumor by binding to tumor-associated antigen of molecular weight (72 kD) identical to that found on 791T cells, the original immunogen.

D. Antibody Binding to Normal Human Cells

1. Reactivity of XMMCO-791 Monoclonal Antibody with Human Peripheral Blood Mononuclear Cells (PBMC) Analyzed by Flow Cytometry

a. Lymphocytes

Using direct and indirect immunofluorescence assays, XMMCO-791 antibody was shown to react strongly with phytohemagglutin (PHA) (Sigma) stimulated lymphoctyes from peripheral blood and negligibly with resting lymphocytes. Maximum binding was observed after 3 days culture with PHA (5 .mu.g/ml), coinciding with maximum DNA synthesis. Comparable results were obtained using concanavalin A (Con A) (10 .mu.g/ml) as mitogen (Table X).

XMMCO-791 antibody precipitated a protein of apparent molecular weight of 72,000 from cells cultured 3 days with PHA. No other bands at 72,000 were precipiated from other cultures of PBMC and no 72,000 molecular weight material was precipitated from PHA-stimulated PBMC using normal mouse immunoglobulin, rather than XMMCO-791 antibody.

b. Blood Granulocytes

Binding of XMMCO-791 to peripheral blood granulocytes has been determined using an indirect flow cytometry assay.

Table XI summarizes a series of tests of 9 lots of XMMCO-791 before and after ultracentrifugation with blood granulocytes from 2 donors. The binding of XMMCO-791 to granulocytes was compared with that of supernatant made by the parental myeloma line, P3/NSI/1-Ag4-1. 8/9 lots showed negative binding. One lot showed moderate binding which became negative after centrifugation.

E. Binding to Other Normal Human Tissues Measured By Immunoperoxidse

XMMCO-791 was extensively evaluated for crossreactivity with normal tissues using the indirect immunoperoxidase conjugate technique. As a control, tissues were simultaneously tested with a purified murine MoAb in the same concentration. The control immunoglobulins used for these purposes were an IgG2 mouse myeloma protein (UPC 10, Litton Bionetics, Kensington, Md.) or a MoAb with reactivity to sheep red blood cells (SRBC).

Fresh tissues slices (2-3 mm thickness and up to 2.0 cm in width) was generously coated with Optimal Cutting Temperature Compound (OCT), (Ames Co., Elkhar, Ind.) and wrapped in aluminum foil. The specimen was "snap" frozen in liquid nitrogen-cooled isopentane (2-methylbutane, VWR, Norwald, Calif.) at controlled temperature (-120.degree. to -150.degree. C.) for 10-15 seconds. The tissue was kept frozen for long-term storage at -70.degree. C., or for short-term storage at -20.degree. C., in a tightly sealed container to prevent evaporation and drying of tissues.

Frozen tissues were allowed to equilibrate slowly overnight by placing in a -35.degree. C. cryostat prior to cutting the frozen sections. Cut sections were dried overnight at room temperature prior to staining.

A circle was etched on each slide around the tissue using a diamond pencil. Slides were then fixed in acetone for 1 minute, followed by a 10 minute PBS wash. Excess fluid was removed and the primary antibody applied for one hour while making sure that the tissue was covered and not allowed to dry out, followed by a 10 minute PBS wash.

Excess fluid was removed and peroxidase-conjugated goat anti-mouse IgG (Tago/product code 6450) (dilute 1:10 in PBS) applied for 30 minutes, followed by a 10 minute PBS wash.

The slides were developed in aminoethylcarbazole (AEC), (0.5 ml AEC stock 120 mg AEC in 15 ml dimethylformamide), 9.5 ml acetate buffer (79 ml of 0.1 M sodium acetate, 21 ml of 0.1 N acetic acid), and 0.05 ml 3% H.sub.2 O.sub.2) for 5 minutes, followed by a 5 minute H.sub.2 O wash. The slides were then counterstained with fresh Mayer's Hemtoxylin for 5 minutes, followed by a 5 minute wash in running tap water and coverslipped with glycerine mounting media. The following is a summary of the results:

Throughout the study, connective tissue (ground substance) was positive.

The vascular endothelium of capillaries and venules in varying tissues were mildly positive but not all venules were positive in any tissue.

Lung tissue from 2/4 autopsy specimens showed some reactivity with about 10% of the alveolar epithelium/macrophages in each of the two specimens; most reactivity was in macrophages.

Kidney tissue showed faint glomeruli staining in 10% of the sample in 1/4 autopsy specimens.

Skin showed staining in the lumen of ducts in 2/4 autopsy specimens.

Colon showed slight staining of glandular epithelium in both colon specimens tested.

Fetal tissue showed some staining in alveolar and bronchial epithelium of the lung; some positive tubule lumen in the kidney;

There was no staining (except for connective tissue and venules) in:

Note: Granulocytes will appear positive in both control and test tissue because of endogenous perxidase.

F. Identification of the Cell Antigen Binding XMMCO-791

1. Osteogenic Sarcoma Derived 791T Cells

791T cells were surface-labeled with .sup.125 I by lactoperoxidase catalyzed radioiodination and lysed. A lysate containing .sup.125 I-labeled membrane protein was obtained. Sodium dodecyl sulfate (SDS) and XMMCO-791 MoAb were added to aliquots of this cell lysate, followed by Sepharose-Protein A beads. The beads were washed and final pellets were suspended in reducing gel sample buffer. After incubation, supernatants were analyzed by polyacrylamide gel electrophoresis. Gels were stained for protein with Coomassie blue and, after drying, autoradiographed.

The antigen precipitated by the MoAb XMMCO-791 from detergent lysates of .sup.125 I-labeled 791T cells migrated in SDS-PAGE slab gels with an apparent molecular weight of 72,000. Both reduced and non-reduced samples gave equivalent results, indicating that the antigen is a monomeric cell surface protein. The relatively broad band suggests that the antigen may be glycosylated, with microheterogeneity in the saccharide residues producing the wide band. When immunoprecipitates were treated with neuraminidase, the apparent molecular weight was reduced to 55,000, indicative of extensive glycosylation. Since microheterogeneity in the band of precipitated material was lost following neuraminidase treatment, this indicated that microheterogeneity was attributable to variations in sialic acid content.

That the gp72 antigen is glycosylated was confirmed by labeling growing tumor cells with .sup.3 H glucosamine and again precipitating a 72,000 molecular weight antigen which was radiolabeled.

2. Other Human Tumor Cell Lines

Seven cell lines known to express the antigen recognized by XMMCO-791 were selected. Four of these were osteogenic sarcomas, two were colon carcinomas, and one was a prostate carcinoma cell line. Negative controls included lung carcinoma, breast carcinoma, and malignant melanoma cell lines.

These 10 human tumor cell lines were surfacelabeled with .sup.125 I, lysed with detergent and subjected to radioimmunoprecipitation using either the XMMCO-791 monoclonal antibody or normal mouse immunoglobulin. The immune precipitates were isolated using SepharoseProtein A and analyzed by SDS-PAGE and autoradiography. The XMMCO-791 antibody precipitated a protein from 791T cells of apparent molecular weight of 72,000 in accordance with previous findings. Five other cell lines contained protein material of apparent molecular weight of about 72,000 that was also precipitated by the XMMCO791 antibody. Colon carcinoma cell line HT29 gave anomalous results, since despite repeated testing, no protein was precipitable with the XMMCO-791 antibody. The cell lines A549, B55, and RPMI 1566, which are unreactive with the XMMCO-791 antibody, failed to express a 72,000 dalton antigen in these immunoprecipitation studies.

It is evident from these studies that the precipitates from the other cell lines which bound XMMCO-791 MoAb were not identical to that from the 791T cell line, although proteins of 72,000 daltons were detected in each case. These differences may be due to antigen heterogeneity (characteristic of carbohydrate-rich glycoprotein antigens) and/or variations in the specific activity of the antigens with respect to radiolabeling. The cell line EB33 clearly showed two distinguishable components of 72,000 and 55,000 daltons that were precipitated by the XMMCO-791 antibody.

a. Effect of Neuraminidase Treatment

Neuraminidase reduced the apparent molecular weight of the precipitate antigens in five of the six cell lines tested to a uniform value of 55,000 daltons. These results indicate, first, that the precipitated labeled protein antigens are glycosylated and, second, that their molecular heterogeneity is attributable to differences in sialic acid content. The protein precipitated from the colon carcinoma cell line HcLo was apparently not affected by treatment with neuraminidase.

b. Effect of Chymotrypsin Treatment

Radiolabeled immune precipitates were treated with chymotrypsin and analysed by SDS-PAGE and autoradiography to compare the XMMCO-791 defined antigens with respect to their susceptibility to proteolysis. The XMMCO-791-defined antigen proved to be very resistant to degradation, the major chymotryptic product in each case being a polypeptide with an apparent molecular weight of 47,000. This would indicate that each of the antigens immunoprecipitated with the XMMCO-791 antibody have at least one chymotryptic site in common. The major 47,000 molecular weight polypeptides from the six tumor cell lines examined did not display the molecular heterogeneity associated with the non-enzyme-treated XMMCO-791-defined antigens. This would imply that the portion(s) of the molecule exhibiting a variable content of sialic acids was (were) removed by proteolysis with chymotrypsin. It may be noted that the antigen from the cell line HcLo was not anomalous with respect to its susceptibility to chymotrypsin. This contrasts with its apparent resistance to neuraminidase since the major polypeptide generated was of 47,000 dalton molecular weight.

3. Effect of Papain on the gp72 antigen

Previous studies have demonstrated that 791T cells treated with papain no longer bind XMMCO-791 antibody, indicative of the degradation or release of XMMCO-791-defined antigen. To define the action of papain more precisely. 791T cells were radiolabeled with .sup.125 I and treated with papain. After centrifugation, the solubilized extract of cell membrane proteins and glycoproteins was subjected to radioimmunoprecipitation using XMMCO-791 antibody and Sepharose-Protein A, and the immune precipitates wer analyzed by SDS-PAGE and autoradiography. The results indicate that papain digestion released an antigenically active protein fragment of apparent molecular weight of 53,000. Neuraminidase treatment of the immune precipitate prepared as above effected a further reduction in apparent molecular weight to 50,000, indicating that this papain fragment was glycosylated and carried sialic acid residues.

4. Purification of the .sup.125 I Labeled gp72 Antigen

To develop a purification procedure for the gp72 antigen, 791T cells were surface labeled with .sup.125 I and lysed with detergent. After centrifugation and dialysis, the soluble extract was fractionated by immunoadsorbent chromatography by passage through two nonspecific adsorbent columns (Sepharose 4B and Affi Gel 10-linked mouse IgG) and a specific immunoadsorbent column (Affi Gel 10-linked XMMCO-791 monoclonal antibody) connected in series. The XMMCO-791 antibody immunoadsorbent was eluted and rechromatographed.

The purified antigen consisted of a major band at 72,000 daltons. There was also a minor band of slightly lower molecular weight; it may be due to carbohydrate heterogeneity.

5. Binding of .sup.125 I-Labeled gp72 to Lectins

The binding of .sup.125 I-labeled XMMCO-791-defined antigen to lectins was investigated. After incubation of aliquots of the antigen preparation (1.times.10.sup.4 cpm) with Sepharose-wheat germ agglutinin (WGA), Sepharose-Con A and Sepharose-Lentil lectin, bound lgycoproteins were eluted and analyzed by SDS-PAGE and autoradiography. The material eluted from Sepharose-WGA migrated with an apparent molecular weight of 72,000, indicating the presence of N-acetyl-glucosamine in this glycoprotein. Equivalent material was also bound to Sepharose-Con A, but no bands were seen with eluates from Sepharose-lentil lectin or Sepharose alone. Of the radioactivity recovered in the eluates from these columns, 69% was present in the fraction bound and eluted from Sepharose-WGA, 49% was bound and eluted from Sepharose-Con A, 23% from Sepharose-lentil lectin and 9% from Sepharose alone.

6. Conclusions

XMMCO-791 MoAb reacts with the osteogenic sarcoma cell line 791T used for initial immunization. The XMMCO-791-defined epitope on this cell line is expressed on a protein of apparent molecular weight 72,000. The MoAb also reacts with cell lines other than this one. An investigation was performed to determine whether the epitope occurred on similar molecules on these other cell lines. It was concluded that molecules expressing this epitope do exist on other tumor cell lines. These molecules are glycoproteins that have different molecular weights. This difference is attributable to varying degrees of sialyation. There are no apparent differences in the polypeptide "backbone" of these antigenic molecules. Radiolabeled immunoprecipitates, prepared with the XMMCO-791 antibody, from three osteogenic sarcoma cell lines (2 OS, 788T, and 278T), the prostate carcinoma EB33 and the colon carcinoma HcLo each contained a protein with a molecular weight of 72,000 as the major constituent and, in some cases, material of lower weight. By neuraminidase treatment of the immune precipitates, this heterogeneity was shown to be due to variations in sialic acid content of the antigens since; in five of the six cell lines tested, such treatment produced homogeneous material of apparent molecular weight 55,000. Chymotrypsin treatment of the immune precipitates produced in each instance a major polypeptide of molecular weight 47,000 which displayed no microheterogeneity. Immunoadsorbent-purified antigen from 791T cells was shown to bind strongly to Sepharose-concanavalin A, confirming the glycoprotein nature of this antigen. These results are summarized in Table XII.

G. Biological Properties of XMMCO-791

1. Complement Dependent Cytotoxicity

Complement dependent cytotoxicity of monoclonal antibody XMMCO-791 was assayed in a 2 hour .sup.51 Chromium release assay. It was found that only rabbit serum served as an effective complement source for lysis of target cells (Table XIII).

2. Antibody Dependent Cellular Cytotoxicity (ADCC)

Experiments were carried out to determine whether XMMCO-791 mediated ADCC reactions (Flannery et al., Eur. J. Cancer Clin. Oncol. (1984) 20:791-98). .sup.51 Cr-labeled target cells (791T, and the human erythroleukemic cell line K562) were incubated with appropriate dilutions of antibody and aliquoted into microtiter wells. Appropriate numbers of human mononuclear cells were added.

XMMCO-791 is not able to mediate ADCC against either 791T cells or K562 cells at concentrations of antibody ranging from 10 to 0.1 .mu.g/ml and with effector:target cell ratios ranging from 0.8:1 to 3:1.

3. In Vivo Inhibition of Tumor Growth

The effect of XMMCO-791 monoclonal antibody alone on the in vivo growth of human tumors was assessed using xenograft systems. 791T cells were implanted subcutaneously into mice immunodeprived by thymectomy and whole body irradiation. Mice were treated with XMMCO-791 antibody starting 1 to 5 days after 791T implantation and were treated at intervals thereafter.

Three injections of either 40 .mu.g or 80 .mu.g did not modify growth of 791T xenografts. In further tests, doses of up to 181 .mu.g/kg body weight (given in 9 divided doses) have not suppressed growth of 791T xenografts.

H. Conjunction of XMMCO-791 With Ricin Toxin A Chain (RTA)

The conjugation of RTA to an immunoglobulin has been thoroughly described in U.S. Pat. No. 4,590,071, the disclosures of which are hereby incorporated by reference. U.S. Pat. No. 4,590,071 fully discloses the technique for RTA purification employing anti-ricin B chain antibodies as well as the conjugation methodology.

I. Characterization of the Immunotoxin

XMMCO-791-RTA was characterized biochemically by SDS-PAGE and functionally by flow cytometry measurement of binding to colorectal carcinoma derived C170 cells and to control MOLT-4 cells (A.T.C.C. Accession No. CRL 1582). It was also characterized by assays that either measured its ability to inhibit protein synthesis in a cell-free system or to kill 791T cells.

In order to determine lot to lot variation, three replicate research lots of immunotoxin were made from the same lot of XMMCO-791 and the same lot of RTA. The conjugations and purifications were done separately but simultaneously. These lots, 51016 A, B, and C, are compared within each section to show lot to lot variation and preliminary data on stability. They have been analyzed by SDS-PAGE, flow cytometry, reticulocyte lysate and whole cell cytotoxicity assays at various times after preparation. The data indicate that the lot to lot variation is minimal, and that the immunotoxins are stable for a least three months.

Results from the assays on a production lot of immunotoxin are also presented in each section. Research lots are prepared in small quantities, whereas production lots are prepared in larger batches under aseptic clean room conditions. The conjugation procedure is the same, but scale up production often produces an increased degree of conjugation. This can be decreased by adjustment of the N-succinimidyl 3-(2-pyridyldithio) propionate (SPDP) ratio. All lots were very similar, both functionally and biochemically.

1. SDS-PAGE Analysis

a. The Assay

SDS-PAGE analysis of immunotoxins run under nonreducing conditions provides information about the amount of free (unconjugated) antibody remaining in the preparation, the means RTA/MoAb ratio, the presence of unbound RTA in the preparation, and the number of molecular species present. Two gel analyses are run on each immunotoxin preparation. A 10% nonreduced SDS-PAGE was run to determine the presence of free RTA, producing poor discrimination between the free antibody and the RTA-MoAb conjugates that tend to cluster near the top of the gel. However, it allows excellent definition of any free RTA, which when present is seen as two bands at 30 kD and 33 kD. Free RTA generally constitutes less than 5% of the total protein in freshly prepared conjugate and RTA bands are rarely seen.

A nonreduced SDS-PAGE was run in a 3-7.5% gradient. This allows accurate detection of the molecular species of conjugate present. Free antibody appears as a band at approximately 160 kD, and immunotoxins with 1, 2, 3 or more RTAs per MoAb appear at appropriate locations reflecting their increased molecular weight.

b. Results

SDS-PAGE analyses were performed on the three replicate research lots of XMMCO-791-RTA and on one of the production lots within two weeks of immunotoxin preparation. The amount of free antibody is similar within the replicate lots, ranging from 1 to 2%. This is typical of free antibody values seen in other research lots. Table XIV summarizes the biochemical characteristics of the replicate and production lots. The RTA to IgG ratio ranged from 3.0 to 3.4 on the replicate lots. Molecular species corresponding to free MoAb or 1-6 RTA/MoAb were seen; this was similar for all research lots. By nonreduced 10% SDS-PAGE, about 5% free RTA was detected.

The production lot, 51203, was used for stability testing, and the results of these tests at one and two months are shown in Table XV, indicating that the product is stable. Four molecular species were detected on nonreduced gradient gels, corresponding to dimers, trimers, tetramers and pentamers. There was no free antibody and no band corresponding to monomers. Nonreduced 10% gels showed that minimal free RTA was present.

c. Summary

Both research and production lots showed little free RTA when freshly prepared. The research lots had up to 1-3% free antibody and molecular species corresponding to 1-6 RTA/MoAb, whereas the production lot had no detectable free antibody and no detectable monomers. Molecular species of up to five RTA/MoAb were present in the production lot. Little deterioration was noted over 3 months.

2. Flow Cytometry Studies, XMMCO-791-RTA Binding to C170 Cell Line

a. The Assay

The competitive inhibition assay, by which flow cytometry is utilized to determine binding of unlabeled XMMCO-791 to C170 colon carcinoma cells, is known to those skilled in the art. This assay is particularly well-suited to assess retention of antibody binding in drug antibody conjugates such as immunotoxins because it measures the ability of unlabeled antibody to specifically compete with FITC-labelled homologous antibody for antigen binding sites.

The reference curve was generated by mixing unlabeled XMMCO-791 standard with FITC-XMMCO-791 and reacting it with C170 cells. This was the positive control. Negative controls included an unlabeled antibody that does not bind to C170 and thus cannot compete. The slope of this line is zero.

The immunotoxin was evaluated by comparing it with its "parent lot"--the MoAb from the same lot used for RTA conjugation. Thus, two sets of mixtures were prepared, the first being a mixture of MoAb from the immunotoxin parent lot with the standard FITC-MoAb and the second, a mixture of the immunotoxin with the FITC-MoAb. When the mixtures were added to C170 cells and assessed for fluorescence intensity by flow cytometry, two lines were generated. Comparison of the slope of the line generated by the parent MoAb with that of the immunotoxin describes the degree of activity retained by the immunotoxin. Since the assay is highly concentration dependent, it was necessary to correct the concentration of the immunotoxin so that is was expressed in terms of antibody concentration instead of immunotoxin concentration.

b. Results

Both the replicate lots had RTA to MoAb ratios of 3-3.4 RTA:MoAb; the more highly conjugated production lot had 4.2 The lots tested for binding at various times after conjugation. The results are shown in Table XVI, and are expressed either as amount of unlabeled antibody necessary to decrease fluorescence by 50% (IF.sub.50) or as percent decrease in line slope. By the latter parameter, it can be seen that there was little difference in the replicate lots, all of which bound about 80% of their parent lot. Moreover, there was little noticeable deterioration, and testing three months after production still revealed binding that was greater than 80% of the parent lot. Table XVI also shows the percent binding of the production lot 51203, which had a higher RTA:MoAb ratio than the smaller lots and bound 48% of the parent lot. There was no decrease in binding two months after conjugation.

c. Summary

Replicate lots made under the same conditions showed uniform binding and little loss of activity compared to unconjugated parent lot. There was no apparent decrease in binding over a three-month period. This demonstrates that RTA:MoAb ratios of 3:1 produce little loss in activity.

The first production lot had an RTA:MoAb ratio of 4.2:1 and decreased binding to about 50% of the unconjugated parent lot. Although this had little effect on either the whole cell cytotoxicity assay, the xenograft therapy trials, or the toxicity, we feel that this represents the upper limit of desirable conjugation. In spite of the higher degree of conjugation, there was no apparent decrease in binding over a two-month period.

3. Reticulocyte Lysate Assay

a. The Assay

This assay measured the ability of XMMCO-791RTA to inhibit protein synthesis in a cell free system, and compared it with the inhibition producted by free RTA. Although both bound and free RTA should theoretically produce the same amount of inhibition, there is usually a slight loss of activity in the conjugate when the two compounds are compared on the basis of RTA molarity.

Rabbit reticulocyte lysate is prepared by lysing rabbit reticulocytes (precursor red blood cells) in water. A nutrient solution is prepared consisting of standard preparations of all amino acids (except leucine), .sup.14 C-leucine, salts, EDTA, hemin (enzyme cofactor), and creatine phosphokinase (the enzyme which catalyzes translation). This solution contains all materials necessary for protein synthesis.

Dilutions of RTA or RTA-containing products were added to types containing the nutrient solution and the lysate. The incorporation of .sup.14C-leucine into protein by the lysate was measured to determine the degree of inhibition of protein synthesis. Results are expressed as a percentage compared to non-RTA containing tubes (100% reference).

Values are expressed as gm/ml producing 50% inhibition of protein synthesis (IC.sub.50). The assay may vary by .+-. 2.5 fold in replicate analysis.

b. Results

Table XVII shows assays on the three replicate immunotoxin lots and one production lot. When plotted in terms of gm/ml, the IC.sub.50 values of the replicate lots ranged between 2.55 and 3.8.times.10.sup.-9. Free RTA gave an IC.sub.50 of about 3.times.10.sup.-10 gm/ml. Within the limits of the assay, the lots with RTA:MoAb conjugation ratios of 3-3.4 gave the same IC.sub.50 values. The IC.sub.50 of the lot with the higher RTA:MoAb conjugation ratio was lower, although still within the range of five fold.

In order to determine if RTA activity was lost after conjugation to MoAb, these same values were expressed in terms of equivalent RTA molarity. Some loss of RTA activity with conjugation was seen in the lots with RTA:MoAb ration of 3, and more loss was noted in the lot with a RTA:MoAb ratio of 4. Other immunotoxins tested show no loss of RTA activity at conjugation ratios of 3.0. This suggest that the conjugated RTA remains biochemically active, but with higher conjugation there is steric hindrance of its activity.

c. Summary

The RLA assay demonstrates that little loss in activity of RTA is seen after conjugation. Since the production lot of immunotoxin had a higher conjugation ratio than the research lots, the IC.sub.50 in terms of gm/ml was different. When expressed in terms of RTA molarity, however, the values are similar, although the lots with lower conjugation ratios have higher relative activity.

The three replicate lots were reanalyzed three months after preparation and the results are shown in the following Table. Little loss of RTA activity is seen.

4. Whole Cell Cytotoxicity Assay

a. The Assay

This assay uses 791T as a positive cell line and MOLT-4 (A.T.C.C. Accession No. CRL 1582) as a negative cell line. It was designed to assess the ability of XMMCO-791-RTA to specifically kill cells expressing the gp72 antigen and is expressed as concentration of immunotoxin producing 50% inhibition of thymidine uptake. MOLT-4, a malignant T cell line lacking the gp72 antigen, is used as the negative cell line.

There are several tumor derived cell lines that carry the gp72 antigen and are specifically killed by XMMCO-791-RTA, as determined using various assays, all of which measure decreased protein synthesis (.sup.75 Se-Selenomethionine) or cell division (thymidine incorporation). Cell lines carrying the gp72 antigen include the colon carcinoma derived cell line C170, used in the flow cytometry analysis of various immunotoxin lots, and colon carcinoma derived cell line LS174.

791T cells or MOLT-4 cells were grown in serum-free medium for 4 to 7 days in T75 tissue culture flasks. 791T, an adherent cell line, was harvested using trypsin-EDTA (Gibco Labs) because it leaves the gp72 antigen intact. MOLT-4 is a suspension cell line. Cells were adjusted to the appropriate concentration, and the concentration used was related to the cells' rate of growth.

For convenience, the assay was carried out on cells in suspension. To prevent 791T cells from adhering during the assay, cells were plated into non-tissue culture coated NUNC 96-well round bottom plates, 100.mu.l/well to a final concentration of 2.times.10.sup.4 cells/well (791T) or 4.times.10.sup.4 cells/well (MOLT-4).

Immunotoxin dilutions were prepared using test tubes precoated with BSA. One hundred .mu.l of each dilution were added to four wells containing cells, for final concentrations of 2.5 to 250,000 ng/ml. The RTA was diluted in a similar fashion to yield final concentrations of 500 .mu.g/mg to 50.mu.g/ml. After a 48-hour incubation at 37.degree. C. and 5% CO.sub.2, 1 .mu.Ci of .sup.3 H-Thymidine was added to each well. After another 24-hour incubation, cells were harvested on filter pads and counted.

b. Results

Each assay of a lot included testing of the immunotoxin against 791T and MOLT-4 cells. Additionally, free RTA was sometimes tested against 791T cells. When the IC.sub.50 values of immunotoxin or free RTA against 791T cells were compared in terms of RTA molarity, the IC.sub.50 of the immunotoxin was more than three logs (Log.sub.10) that of the free RTA. Also, there was a 3-log.sub.10 difference between IC.sub.50 of XMMCO-791-RTA immunotoxin against 791T and MOLT-4 cells. The IC.sub.50 of XMMCO-791-RTA on negative control MOLT-4 cells (in RTA equivalents) was very similar to that of free RTA on 791T cells.

Table XIX summarizes the IC.sub.50 of the replicate and production lots in terms of nanograms/ml and molarity of RTA. Although the IC.sub.50 values were quite similar, there was a loss of potency as the conjugation ratio increased, reflected by a higher IC.sub.50.

c. Summary

The whole cell kill assay measures the potency of intact immunotoxin. There was minimal lot to lot variability, as shown by the results with the three replicate lots. Even though the production lot had a decrease in binding due to the higher conjugation ratio, the IC.sub.50 was still similar to those of the replicate lots. This may mean the decreased binding was somewhat compensated for by the increased number of RTA molecules. These IC.sub.50 values were similar to those seen with other immunotoxins.

J. XMMCO-791-RTA Therapy of Human Tumor Xenografts

This study was performed to determine the therapeutic efficacy of XMMCO-791-RTA against human tumor xenografts expressing the gp72 antigen that is recognized by the XMMCO-791 antibody.

1. Procedure

Nude mice were implanted with tumor derived from an established xenograft according. (minced to small 1mm pieces) Three to four days after tumor implantation, groups of up to 10 mice received immunotoxin by intraperitoneal injection and their tumor growth was compared with that in controls receiving saline. A range of treatment schedules, including daily injections for 20 days and a protocol involving treatment 3 times per weeks were employed.

2. Results

In all, three lots of immunotoxins were evaluated, Lots 41203B, 50903 and 51203.

There were three xenograft tests using two different conjugates. In the first series, XMMCO-791-RTA (Lot 41203B) was tested against osteogenic sarcoma 791T xenografts using various treatment protocols. One test scheduled treatment 3 times weekly at a dose of 4.8 mg/kg intraperitoneally for nine doses (total dose 43 mg/kg). Compared with the control group, this produced a significant inhibition of tumor growth (FIG. 1). The most effective response with this preparation was achieved in mice receiving eight doses of 12 mg/kg (total dose 96 mg/kg), administered at 2 to 3 day intervals. This is illustrated in FIG. 1c, showing that during the treatment there was a marked inhibition of tumor growth. As expected, tumor growth was observed in the treated group after cessation of treatment. In this respect, the rate of tumor growth in the treated group was comparable to that in the controls, although it is displaced in terms of tumor size. FIG. 1d summarizes these trials and shows a good correlation between total dose of XMMCO-791-RTA and therapeutic response, even though different dosing schedules were used.

The overall toxicities observed with this immunotoxin are summarized in FIG. 1e. As illustrated, there was no mortality with doses up to 100 mg/kg. The only observed effect was a gradual weight loss as doses escalated from 40 mg/kg to 100 mg/kg.

In the second series of trials using XMMCO-791-RTA (Lot 50903), the treatment schedule was 5 mg/kg daily for 20 days (total dose 100 mg/kg). This protocol produced marked inhibition in growth of sarcoma derived 791T xenografts (FIG. 2a) and colon carcinoma derived C170 xenografts (FIG. 2b). There was significant toxocity, however, with 3/10 mice dying in both tests.

Taking into consideration the experience gained with two immunotoxin preparations, a third series of test were carried out using XMMCO-791-RTA (Lot 51203). The immunotoxin was given in 20 daily doses of 3 mg/kg/day (total dose 60 mg/kg); the response elicited against colon carcinoma C170 xenografts demonstrated significant suppression of tumor growth. This is illustrated in FIG. 3 which shows growth of colon carcinoma C170 xenografts in immunotoxin-treated mice compared with controls. Furthermore, tumor weights in the immunotoxin treated group were significantly less than those in controls (T/C=0.52 p<0.02) at the completion of the trial.

A. Production of Hybridomas

Balb/c mice (Bantin & Kingman, U.K.) were immunized with carcinoembryonic antigen (CEA) derived from a perchlorate extract of a colon carcinoma liver metastasis. The immunization schedule consisted of 10 .mu.g CEA in complete Freund's adjuvant (CFA) given intraperitoneally on Days 0 and 7, and 20 .mu.g of CEA (Also in CFA) given intraperitoneally on Days 25, 56 and 63.

Three days after the last antigen boost, spleen cells from an immunized mouse were aseptically removed. Following procedures as outlined elsewhere (Galfre et al (1977) Nature 266:550, which is incorporated by reference), 10.sup.6 spleen cells were fused with 10.sup.6 cells of P3-NSI-Ag4-1 (A.T.C.C. Accession No. TIB18), a hypoxanthine-methotrexate-thymidine (HMT) sensitive murine myeloma cell line. Using polyethylene glycol (PEG), hybrid cells were placed into 96-well culture plates, (Costar, Cambridge, Mass. #3596) on medium containing a feeder layer of rat peritoneal exudate cells (2.5.times.10.sup.3 cells/well). Cells were cultured in Dulbecco's Modified Eagle's Medium (Flow Labs, Irvine, U.K.) containing 15% Fetal Calf Serum (Myoclone, Gibco, Paisley, U.K.) and hypoxanthine (10.sup.-4 M), thymidine (1.6.times.10.sup.-5 M) (both from Sigma, Dorset, U.K.) and methotrexate (10.sup.-5 M) (Lederle, Hampshire, U.K.).

Within two weeks post fusion, cultures of hybridoma cells were tested for antibody binding to CEA by enzyme immunoassay (EIA). Cultures that were positive were cloned using limiting dilution, plating 1-3 cells/well into 96 well culture plates. Wells containing only one colony were identified by microscopic examination, then tested for reactivity with CEA and normal colon antigen (NCA) by EIA and radioimmunoassay (RIA), and for reactivity on extranuclear membranes from primary colorectal tumors and normal colon by EIA. The clone designated XMMCO-228 was found to stably secrete immunoglobulin which was determined to be of the IgG2a subclass by solid phase RIA using standard methods. Hybridoma XMMCO-228 is presently on deposit with the American Type Culture Collection (A.T.C.C.), 12301 Parklawn Dr., Rockville, Md. 20852, U.S.A. The deposit was made on August 14, 1986, and given A.T.C.C. Accession No. HB 9174.

Balb/c mice (Bantin & Kingman, U.K.), 6-10 weeks old, were used to culture the hybridoma peritoneally. Approximately 10.sup.7 hybridoma cells were injected into mice that had been pretreated 3 weeks earlier with 0.5 mls of pristane (Aldridge, Gillingham, Dorset, U.K.) injected intraperitoneally (i.p.). The resultant ascites fluid, collected 3 weeks after injection of the hybridomas, contained on average 5 mg/ml of the antibody as determined by measuring immunoglobulin level according to the method of Price and Baldwin ICRS Med. Sci. (1984) 12:1000-01, which is incorporated by reference.

The antibody in ascites fluid was purified by affinity chromatography using a Sepharose - Protein A column using methods well known by those skilled in the art.

The hybridoma was grown in vitro in Dulbecco's Minimal Essential Medium (Flow Labs, Irvine, U.K.) with 10% Fetal Calf Serum (Myoclone, Gibco, Paisley, U.K.) and hypoxanthine (10.sup.-4 M), thymidine (1.6.times.10.sup.-5 M) (both from Sigma, Dorset, U.K.) and methotrexate (10.sup.-5 M) (Lederle, Hampshire, U.K.) in plastic 300 ml bottles. Cell concentration was 10.sup.5 cells/ml medium over a culturing period of 4-5 days, with a MoAb concentration of 4 .mu.g/ml medium, and a doubling time of 12 hours.

B. In Vitro Antibody Binding to Cell Lines Measured by Flow Cytometry

The binding of XMMCO-228 and XMMCO-791 MoAbs to MKN45 cells was determined by flow cytometry employing methods well known to those skilled in the art and described above. The gastric carcinoma cell line MKN45 expresses both the 72 kD antigen with which the XMMCO-791 MoAb reacts and the CEA antigen with which XMMCO-228 reacts. B14/B8 is a MoAb directed against breast cancer, which reacts with the normal cross-reacting antigen (NCA) subsection of the CEA antigen. Normal mouse serum was used as a control. The results are summarized in Table XX.

C. In Vivo Antibody Binding to Normal Tissues Measured by Immunoperoxidase

The technique for Indirect Immunoperoxidase staining has been described above. Table XXI is a summary of the results obtained with XMMCO-228 MoAbs in humans:

It is significant that XMMCO-228 MoAbs do not bind to granulocytes. Normal colon antigen (NCA) appears on such diverse tissues as normal colon, granulocytes, and progenitor bone marrow cells. Because the CEA antigen has a NCA subsection, many anti-CEA MoAbs also bind the NCA on various normal tissues, which XMMCO-228 MoAbs do not.

D. XMMCO-228 Ricin Toxin A Chain (RTA) Conjugates

As in the case of XMMCO-791-RTA conjugates described above, the conjugation technique, including purification of the A chain of ricin, is disclosed in U.S. Pat. No. 4,590,071, the disclosures of which are hereby incorporated by reference.

E. XMMCO-228-RTA In Vitro Cytotoxicity

In vitro cytotoxicity of XMMCO-228-RTA conjugates was determined by the .sup.75 Se-Selenomethionine Assay. Single cell suspensions of gastric carcinoma derived MKN45 cells were inoculated with immunotoxin or RTA for 15 minutes, washed and added to flat bottomed microtiter plates at a concentration of 10.sup.5 cells/0.2 ml of immunotoxin or RTA in RPMI medium containing 10% fetal calf serum. Various amounts of immunotoxin were added to each well. Twenty-four hours later, .sup.75 Se-Selenomethionine was added to each well. Cultures were further incubated for 16 hours, washed, dried on the bottom of the wells, and the wells were counted in a gamma counter. The results are summarized in Table XXII, expressed in terms of 50% inhibition concentration.

These data indicate that when XMMCO-791-RTA and XMMCO-228-RTA are administered in a "cocktail" form, the cytotoxicity is at least additive, and possibly synergistic. They also show the efficacy of immunotoxin (or antibody directed RTA) compared to RTA alone.

F. XMMCO-228-RTA Therapy of Human Tumor Xenografts

For each data point, 7-10 nude mice were implanted with LS174T cells, a human colon adenocarcinoma derived cell line (A.T.C.C. Accession No. CL 188). Three days post-implant, the first dose of the treatment compound was administered, followed by 10-20 doses until the schedule dose was completed. The results are summarized in Table XXIII, expressed as a ration of the tumor weight in treated mice divided by the tumor weight in control (untreated) mice. The results, in terms of mean tumor diameter, are summarized graphically in FIG. 4.

These results demonstrate the efficacy of XMMCO-228RTA in the therapy of human tumor xenografts.

A. XMMCO-791-RTA+XMMCO-228-RTA Conjugate Cocktails

A cocktail of XMMCO-791-RTA and XMMCO-228-RTA consists of a mixture of XMMCO-791-RTA and XMMCO-228-RTA, in any ratio, and administered to a cancer cell host in a dose range of about 0.01 mg/kg/day to 20.0 mg/kg/day. The cocktail is administered perenterally, generally by intravenous infusion or intrapercitoneal injection in a suitable vehicle, such as phosphate buffered saline or the like.

B. In Vitro Cytotoxicity of Conjugate Cocktails

In vitro cytotoxicity of XMMCO-228-RTA conjugates was determined by the .sup.75 Se-Selenomethionine assay. Single cell suspensions of gastric carcinoma derived MKN45 cells were inoculated with immunotoxin or RTA for 15 minutes, washed and added to flat bottomed microtiter plates at a concentration of 10.sup.5 cells/0.2 ml of immunotoxin or RTA in RPMI medium containing 10% fetal calf serum. Various amounts of immunotoxin were added to each well. Twenty-four hours later, .sup.75 Se-Selenomethionine was added to each well. Cultures were further incubated for 16 hours, washed, dried on the bottom of the wells, and the wells were counted in a gamma counter. The results are summarized in Table XXII above, expressed in terms of 50% inhibition concentration.

These data indicate that when XMMCO-791-RTA and XMMCO-228-RTA are administered in a "cocktail" form, the cytotoxicity is at least additive, and possibly synergistic. They also show the efficacy of immunotoxin (or antibody) directed RTA compared to RTA alone.

C. XMMCO-791-RTA+XMMCO-228-RTA Conjugate Cocktail Therapy of Human Tumor Xenografts

For each data point, 7-10 nude mice were implanted with MKN 45 cells, derived cell line. Three days post-implant, the first dose of the treatment compound was administered, followed by 10-20 doses until the scheduled dose was completed. The results are summarized in Table XXIV, expressed as a ratio of the tumor weight in treated mice divided by the tumor weight in control (untreated) mice.

These data indicate that when XMMCO-791-RTA and XMMCO-228-RTA are administered in a "cocktail" form, the cytotoxicity is at least additive, and possibly synergistic. They also show the efficacy of a cocktail consisting of XMMCO-791-RTA and XMMCO-228-RTA in treating cancer cells in a host.

A. Production of XMMBR-B14 Hybridomas

Balb/c mice were immunized with 521AM whole cells derived from ascites breast metastasis. The immunization schedule consisted of 6 intraperitoneal injections of 10.sup.6 cells at weekly intervals. A final intravenous boost of 3.times.10.sup.5 cells was given 3 days prior to fusion.

Three days after the last antigen boost, spleen cells from an immunized mouse were aseptically removed. A single cell suspension was obtained using an 80 mesh wire screen grid #1985-00080 (Bellco). Cells were washed with Iscoves Complete Medium (Grand Island, N.Y., N.Y.) and counted. Sp2/0-Ag14 cells (A.T.C.C. Accession No. CRL 1581), a hypoxanthine-aminopterinthymidine (HAT) sensitive murine myeloma cell line, were washed 3 times and counted. Using polyethylene glycol (PEG 4000 Merck) 30% v/v, 10% DMSO and 60% Iscoves, the two cell types were fused at a ratio of 2 spleen cells per myeloma cell. The fusion products were plated into 96-well culture plates at a concentration of 10.sup.5 myeloma cells per well. Cells were cultured in Iscoves with 20% fetal bovine serum and .beta. mercaptoethanol with HAT medium (hypoxanthine 136 mg/100 ml, aminopterin 0.018 mg/100 ml, thymidine 136 mg/100 ml).

Within two weeks post fusion, cultures of hybridoma cells were tested for antibody binding to 521 AM tumor cell membranes by dotting onto nitrocellulosee membranes at 200 ng using a Vector Labs ABC kit to test for reactivity. Wells that gave a a blue dot were then screened against fibroblast membrane extracts from the same patient. Cultures that were positive were cloned using limiting dilution, plating 1-3 cells/well into 96 well culture plates. Wells containing only one colony were identified by microscopic examination, then tested for reactivity. The clone designated XMMBR-B14 was found to stably secrete immunoglobulin which was determined to be of the IgG1 subclass. Hybridoma XMMBR-B14 is presently on deposit with the American Type Culture Collection (A.T.C.C.), 12301 Parklawn Dr., Rockville, Md. 20852, U.S.A. The deposit was made on Jan. 14, 1987, and given A.T.C.C. Accession No. Hb 9308.

Balb/c mice (Bantin & Kingman, U.K.), 6-10 weeks old, were used to culture the hybridoma peritoneally. Approximately 10.sup.7 hybridoma cells were injected into mice that had been pretreated 3 weeks earlier with 0.5 mls of pristane (Aldridge, Gillingham, Dorset, U.K.) injected intraperitoneally (i.p.). The resultant ascites fluid, collected 3 weeks after injection of the hybridomas, contained on average 5 mg/ml of the antibody as determined by measuring immunoglobulin level according to the method of Price and Baldwin ICRS Med. Sci. (1984) 12:1000-01, which is incorporated by reference.

The antibody in ascites fluid was purified by affinity chromatography using a Sepharose - Protein A column using methods well known by those skilled in the art.

The hybridoma was grown and cloned in vitro in Iscoves Medium with 20% fetal bovine serum and .beta. mercaptoethanol in plastic 300 ml bottles. Cell concentration was 10.sup.5 cells/ml medium over a culturing period of 4-5 days, with a MoAb concentration of 4 .mu.g/ml medium, and a doubling time of 12 hours.

B. In Vitro XMMBR-B14 Binding to Cells Measured by Flow Cytometry

The binding of XMMBR-B14 to cell lines and primary carcinoma-derived cells was determined by flow cytometry employing methods well known by those skilled in the art and described above. The tests (Table XXV) showed that XMMBR-B14 detects an antigen expressed on breast and colon carcinoma cell lines. Further studies using purified protein preparations have established that the epitope defined by XMMBR-B14 is found on an antigen of the classification termed CEA/NCA (carcinoembryonic antigen/normal cross-reacting antigen). This is the portion of the CEA molecule that also cross-reacts with exposed normal cross-reacting antigen. It also reacts with cells derived from two primary colon carcinomas. Normal mouse immunoglobulin (NMIg) or normal mouse serum (NMS) were used as controls.

C. XMMBR-B14 Reactivity With Colon Carcinoma and Normal Colonic Mucosa Measured by Enzyme Immunoassay (EIA)

XMMBR-B14 was tested for reactivity with primary colon carcinoma membrane and normal colonic mucosa using standard EIA methods well known by those skilled in the art. The test, summarized in Table XXVI, shows that XMMBR-B14 reacts with colon carcinoma membrane.

D. XMMBR-B14 Binding to CEA and NCA

The Reactivity of XMMBR-B14 with CEA and NCA preparations was determined by a solid phase radioimmunoassay. Briefly, antigen preparations were coated into wells of microtest plates. Monoclonal antibody was then added, incubated 1 to 2 hours and wells washed. .sup.125 I-labelled (F(ab) 2 fragments of rabbit anti-mouse IgG were then added to detect bound murine MoAb.

In the first test MoAb binding to semi-purified CEA (Rogers) and NCA (B3991) was compared. XMMBRB14 bound to both NCA and CEA (NCA:CEA ratio 1.2:7).

E. XMMBR-B14 Dose Response on XMMCO-228-RTA Cytotoxicity Against MNK45 Cells

Table XXVII shows the increase in cytotoxicity of XMMCO-228-RTA immunotoxin against MNK45 cells when increasing amounts of XMMBR-B14 are added.

Six different curves were generated using concentrations of XMMCO-228-RTA ranging from 1 ng/ml to 100 ng/ml. To these concentrations were added varying amounts of XMMBR-B14 ranging from 0 to 10 .mu.g/ml. The calculated IC.sub.50 values for each curve are shown in Table XXVII.

The present invention provides efficacious, novel compounds and methods for the therapy of various cancer cells. The cytotoxic conjugates are particularly efficacious when administered in cocktail form or in the presence of certain unconjugated monoclonal antibodies.

Although the present invention has been described in some detail by way of illustration and example for purposes of clarity of understanding, it will be obvious that certain changes and modifications may be practiced within the scope of the appended claims.