WO1985002411A1

WO1985002411A1 - Cell line and monoclonal antibody

Info

Publication number: WO1985002411A1
Application number: PCT/AU1984/000237
Authority: WO
Inventors: Ian Farquhar Campbell Mckenzie
Original assignee: The University Of Melbourne
Priority date: 1983-11-25
Filing date: 1984-11-16
Publication date: 1985-06-06
Also published as: IL73620A0; EP0162070A1

Abstract

A cell line which produces a monoclonal antibody having a predominant specificity to breast cancer antigens.

Description

TITLE: CELL LINE AND MONOCLONAL ANTIBODY Background of the Invention This invention relates to a new cell line and to monoclonal antibodies produced thereby. The invention also relates to a method of detecting breast cancer using the antibodies embodying the invention. Antibodies are primary components of immune defence systems and are made by the white blood cells (lymphocytes) in response to the presence of foreign material such as bacteria, virus, tumor cells or even inorganic chemicals. Such foreign material, known as antigens, are chemically highly i de nt i f ia bl e by means of the specific proteins produced thereby. In general, antibodies, which are produced by the lymphocytes in response to the presence of antigens, are Y-shaped protein molecules, each of w h ic h can bind uniquely to the antigen and render same inert. Mammalian antibody defence systems are highly flexible and it has b e e n estimated that the average person can make, a n t i b od ie s to more than a million different foreign molecules. Familiar antibodies take the form of antiserums such as snake or spider venom antiserum. Other examples include anti-tetanus and anti-rabies antiserums, which are used in the treatment of such diseases. These antiserums are traditionally made by injecting forms of t.he various disease-carrying micro-organisms (the antigens) into domestic animals, such as rabbits, sheep, horses and goats. The selective and specific binding capacity of antibodies has made them useful for another purpose - the detection of very small quantities of foreign molecules or antigens in blood, body fluids and tissues. Antibodies have been used to diagnose disease since their discovery in the late 19th century, at first in relatively crude systems, but with increased sophistication until many millions of diagnostic tests are carried out annually at the time of writing. Until recently however the full potential of antibody-based diagnostic systems has not been realized. This has been due mainly to the method of production as each antiserum produced by an animal is a complex mixture of many antibodies, only a tiny fraction of which is relevant to the required purpose. In addition, no two batches of antiserum are every the same as no animal responds to the introduction of antigen in the same way twice. However in 1975, Kohler & Milstein made a breakthrough in the United Kingdom in antibody production, which has had an enormous impact throughout biology and medicine. These scientists discovered a way to make unlimited quantities of identical antibody molecules from a single line or clone of cells in a test tube. These identical molecules were called monoclonal antibodies and it is the use of one of these monoclonal antibodies that is the key to the present invention. Such antibodies exploit the remarkable specificity of the immune system, and bring to fields of medicine and biotechnology a degree of specificity that had previously been impossible to achieve. Monoclonal antibodies are typically produced by using a cultured cancer cell (a myeloma) and a living cell taken from the spleen of a mammal (for example BALB/c strain of mouse). Preferably the spleen cell has already been primed with an antigen and is already producing a specific antibody. The new cell thus formed is called a hybridoma. The hybridoma displays the characteristics of both parent cells, in that it produces the same antibodies produced by the parent spleen cell and also exhibits the same vigorous growth, production and longevity characteristic of a cultured cancer cell. Hybridomas formed by such fusions are cultured and cloned, and clones which produce antibodies demonstrating a specificity for a desired antigen are selected and cultivated. In this way, an ample supply of a monoclonal antibody specific to the antigen used to immunise or "prime" the parent spleen cell can be obtained. The hybridoma is thus a highly sensitive diagnostic agent for the antigen in question. It is anticipated that the excellent specificity and ready availability of large quantities of monoclonal antibodies will cause a revolution in immunodiagnostic testing procedures within a relatively short time.

The ability to obtain pure (monoclonal) antibodies specific to individual antigens has resulted in a proliferation of the use of monoclonal antibodies. Applications include the testing and/or treatment against many antigens derived from bacterial, viruses, hormones, parasites and even various chemicals, although emphasis to date has been placed on the diagnosis of a variety of bacteria and viral diseases, on blood and tissue typing, and on assays for several hormones and enzymes whose presence in blood or urine can itself be used as a diagnostic.

The application of monoclonal antibody techniques to bi otec hnological research has been well described in the literature, including Kohler et al in NATURE, Vol 256,495/497 (1975); Walsh in NATURE Vol 266,495 (1977); R.L. Gatz et al in BIO/TECHNOLOGY 337 of June 1983; L.A. Herzenberg et al in IMMUNOGENETICS 1982; George S. Eisenbarth in ANALYTICAL BIOCHEMISTRY 111, 1/16 (1981); U.S. Patents Nos. 4,172,124 and 4,359,457; and Australian Petty Patents Nos. 529,210; 530,850 and 530,851. Prior to this invention however, it was not known whether specific monoclonal antibodies could be used to test human blood serum or other secretions for breast cancer. Previously it had been the practice to test suspected mammary tissue that had been removed surgically from the patient.

A number of monoclonal antibodies have now been made which react primarily with human carcinoma of the breast, but also react with other tumours and have a variable reaction with normal breast tissues (see for example Foster et al, Virchows Arch (Pathol. Anat.) 1982; Vol 394, pages 279-293; Papsidero et al. Cancer Research 1983; Vol 43 pages 1741-1747; Schlom et al, Proc. Natl. Acad. Sci. 1980; Vol 77, pages 6641-6645; Ceriani et al. Somatic Cell Genetics 1983; Vol 9, pages 415-427; and Thompson et al, J. Natl. Cancer Inst. 1983; Vol 70, pages 409-419). Unfortunately, none appear to be entirely specific for carcinoma of the breast and indeed no monoclonal antibody has been described which has been conclusively proven to be tumour specific for human carcinomas. For diagnostic and possibly therapeutic purposes, it is therefore desirable to have a panel of antibodies which in toto define tissue and/or tumour specificity. Brief Summary of Invention This invention provides a new cell line, which has been designated 3E1-2, which produces a monoclonal antibody having a predominant specificity to breast cancer antigens. The invention also provides a monoclonal antibody having a predominant specificity to breast cancer antigens. The invention also provides a method of detecting breast cancer in blood serum secretions or tissue comprising the step of applying to an immunoassay containing the blood serum, secretion or tissue to be tested a monoclonal antibody having a predominant specificity to breast cancer antigens and detecting the presence or absence of a reaction in said immunoassay. The invention also provides for a continuous cell line identified as clone 3E1-2 which produces breast carcinoma antibodies in vitro in hypoxan th ine-aminopterin-thym idine medium comprising a fused cell hybrid of human adenocarcinoma-primed BALB/c mouse spleen cells, and NS-1 mouse myeloma cells, said antibodies reacting with a breast tissue antigen. Brief Description of the Drawings Figs. 1 and 2 are graphs showing the results of serum inhibition study conducted on a number of patients. General Description of Preferred Embodiments Monoclonal antibody 3E1-2 is produced against human carcinoma of the breast by the well-known method of immunizing BALB/c mice with breast cancer cells and performing a fusion of spleen cells from the immunised mouse with NS-1 myeloma cell line using polyethylene glycol as the fusing agent. Antibody 3E1-2 is of the IgM class and is detected using an immunoperoxidase method to stain cancer of the breast cells. Breast antigens are present in significantly greater than normal amounts in the blood serum, secretions or tissue of patients with carcinoma of the breast. The antibody reacts very strongly with carcinoma of the breast antigens, and weakly or not at all with normal breast tissue. The antibody the subject of this invention is uniformally negative with the surface of sime in vitro cancer cell lines. However, a 100% positive result has been found on breast cancer tissues. Some positive results were also derived from other cancer tissues such as lung or colon; however, the strength of the reaction was substantially less than the results derived from the breast cancer tissue. Positive reactions are obtained on formation fixed breast cancer tissue using the immunoperoxidase method so the antibody provides a useful reagent for histological diagnosis of breast cancer. Antibody 3E1-2 therefore provides a highly satisfactory testing medium on tissue and blood serum for breast cancer using any suitable technique such as ELISA or radioi mmunoassay, both of which are well-known and require no further description. Since the antibody is effective in serum testing, one effect of the method of this invention is that it will be no longer necessary to physically remove tissue from the breast to test for cancer. The test may be carried out on blood serum or other secretions and provides substantial savings in time and cost, as well as providing a simple diagnostic procedure suitable for general practitioners. Further, tests previously produced to test mammary tissue for breast cancer were notoriously unreliable. Lack of sensitivity and specificity often lead to false results, either positive, or regretably negative. The method of the present invention provides results which are substantially more accurate. This is due to the high specificity of the antibodies, produced by the 3E1-2 cell line and extremely high sensitivity of those antibodies to breast carcinoma antigen. As will be apparent from the above, one major feature of the invention lies in that a breast cancer test can be carried out using blood serum or other secretions as well as on tissue surgically removed from a suspect breast so that attractive alternatives to tissue testing are provided.

Detailed Description of Preferred Embodiments Abbreviations Used

ABA Azobenzenearsonate

BSA Bovine Serum Albumin

CEA Carcineombryonic Antigen

CPM Counts per minute

DMEM Dulbecco's Modified Eagles Medium

EDTA Ethylenediamine tetra-acetic acid

FCS Foetal Calf Serum

HAT Hypoxanthine, Aminopterin, Thymidine medium

HCG Human Chorionic Gonaditrophin

HLA Human Leukocyte Antigens

L15 Leibovitz's medium

MFGM Milk Fat Globule Membrane

NCS Newborn Calf Serum

NS-1 P3-NSl-Ag4-1

PBS Phosphate Buffered Saline

PLP Periodate-Lysine-Paraformaldehyde

PVC Poly Vinyl Chloride

RAMG Rabbit Anti Mouse Immunoglobulin

RPMI Roswell Park Memorial Institute medium

SAMG Sheeop Anti Mouse Immunoglobulin

SRBC Sheep Red Blood Cells

TTS Triton X-100, Tris, Saline

Materials and Methods Cell Lines

All human cell lines were maintained in RPMI-1640 medium containing 10% NCS, penicillin, streptomycin and glutamine. The mouse myeloma P3-NSI-Ag4-1 (NS-1) was maintained in DMEM containing 10% NCS, penicillin, streptomycin and glutamine. (All media and additiives were obtained from Flow Laboratories, Sydney, Australia). Adherent cell lines were harvested after the culture medium was decanted, and the cell monolayer exposed to 0.20% trypsin (commonwealth Serum Laboratories, Melbourne, Australia) at 37°C, for two minutes, followed by the addition of NCS to neutralize the trypsin. When the cells detached, they were harvested and washed three times in Leibovitz's medium (L-15) and 0.5% Bovine Serum Albumin (BSA; Miles Laboratories, Ind.) prior to use. To demonstrate that the antigen recognised by 3E1-2 was not trypsin sensitive, tisdsue sections were trated with PBS-versene solution, containing various concentrations of trypsin for up to 30 minutes at 37°C. Fresh Tumour Tissue A breast carcinoma of the ductal type was obtained from a surgical specimen, a single cell suspension produced in serum free DMEM, centrifuged at 350g for five minutes, then resuspended in PBS for immunisation. Fusion and Maintenance of Hybridomas Two (CBA x BALB/c)F₁ mice were immunised 20 days (i.v.) and three days (s.c.) prior to fusion, with 5x10⁶ tumour cells/ml in PBS. On the day of fusion the spleens were removed, processed and fused with the NS-1 myeloma in a manner which will be well-known to persons skilled in the art. After screening for antibody production the hybrids were cloned by the limiting dilution technique. The hybridoma selected for study in this report was also grown in (CBA x BALB/c) F₁ mice injected previously with Freund's adjuvant, from which ascites and/or serum was collected. Serological Methods The sheep anti-mouse immunoglobulin (SAMG) rosetting assay described in The Journal of Immunological Methods 1978 Vol 20 at pages 173 to 183 was performed in 96 well microtiter plates (Disposable Products, Adelaide, Australia) by mixing 25ul target cells (5x10⁶/ml) with culture supernatants or dilutions of antisera (25ul) for 30 minutes at 4°C, by washing three times in L-15-0.5% BSA and then by adding 25ul of SRBC coated with SAMG using chromic chloride. The SRBC-SAMG-target cell suspension was sedimented by centrifugation at 200g for three minutes and the percentage of rosettes formed counted after 30 minutes at 4°C. Preparation of Lymphocytes Heparinised blood was collected and centrifuged at 350g for 10 minutes, and the platelets removed. Then 1ml of Ficoll Paque (Pharmacia, Uppsala, Sweden) was placed under 6mls of blood; centrifugation of the suspension at 750g for 10 minutes left a lymphocyte enriched interface which was separate from the red blood cells and granulocytes. These were collected, washed three times and resuspended in L15-5% BSA. Tissue Sections Normal and neoplastic tissues were obtained immediately after surgery, snap frozen in a liquid nitrogen-isopentane (BDH Chemicals) slurry, and stored at -70°C. Tissue sections (6um) were cut in a cryostat and placed on glass slides previously washed for two hours in 90% alcohol. Sections (5um) were also cut from formalin-fixed, paraffin embedded blocks and dried onto glass slides. The slides were rehydrated in xylene and alcohol, ready for the immunoperoxidase and immunofluorescence assays. Immunoperoxidase Staining After rehydration formalin-fixed tissue sections were treated with 0.5% H₂O₂ in PBS for 30 minutes to remove endogenous peroxidase activity, then washed in PBS for 15 minutes (this step was omitted for fresh cryostat sections). Tissue sections were covered with 1:10 dilution of supernatant containing antibody incubated in a humidified atmosphere for 40 minutes at room temperature, then washed for 15 minutes in PBS and 0.2% gelatin. An ammonium sulphate precipitate of rabbit anti-mouse immunoglobulin (previously absorbed with fresh human spleen to remove non- specific antibody binding) diluted 1:100 was then applied and incubated for 20 minutes. The slides were washed for 15 minutes in PBS and 0.2% gelatin, and then covered with a 1:20 dilution of swine anti-rabbit immunoglobul in antiserum conjugated with horse-radish peroxidase (DAKO- Immunoglobul ins, Copenhagen, Denmark) for 20 minutes. After the slides were washed in PBS and 0.2% gelatin, di aminobenzi di ne (DAB; 1.5mg/ml, Sigma Chemical Co.) was added for 4-5 minutes. The excess DAB was removed and the slides washed in running tap water for 5 minutes, counter stained with haematoxylin and then mounted. A monoclonal antibody directed against the Ly-2.1 antigen of the mouse lymphocyte was used as a non-reactive control and a medium control was also included for each experiment. Immunofluorescence Tissue sections were stained by indirect immunofluorescence. Sections were incubated with a 1:10 dilution of the hybridoma culture supernatant for 40 minutes in a humidified atmosphere at room temperature. The sections were subsequently washed in PBS, and stained with a sheep anti-mouse IgG antiserum conjugated with fluorescein. Stained sections were examined with a Leitz Orthoplan Fluorescene microscope by transmitted darkfield illumination with a narrow band blue excitation. Immunoglobulin Assays The immunoglobulin isotype of the monoclonal antibody 3E1-2 was determined by Ouchterlony gel diffusion using class specific rabbit antisera to IgM, IgGl, IgG2a, IgG2b (Meloy Lab., Springfield, Va., USA). The antibody titre of 3E1-2 was determined by subjectively examining histological sections (known positive sections) stained by antibody in dilution; the titer being recorded as the next and last tissue section which did not stain. Serum Collection and Storage Samples of blood were obtained from patients with carcinoma of the breast, and normal female controls. Serum was separated by allowing the blood to clot, then spinning at 750g for 10 minutes. The serum was removed and stored at -70°C until use. Kidney membrane was used as a source of antigen (as it is eadily available to us than breast tissue), preparations were made by taking 10° cells (cell lines or fresh tissue) and homogenising in 0.25M surcorse (25mM Tris, ImM EDTA, pH7.4) with a Dounce homogeniser at 4°C. The homogenate was spun at 2000g for one minute at 4°C to remove nuclei, and the resulting supernatant spun at 4,200 g for 10 minutes to pellet the crude membrane preparation and the portein content determined by the method described in 'Analytical Biochemistry' 1976 Vol 72 at pages 248 to 254. The assay was performed by adding 50ul of a 200ug/ml stock solution of crude kidney membranes in PBS to a 96 well flexible PVC plate. Membranes were fixed to the plate with 50ul of 0.1% glutaraldehyde at room temperature for 15 minutes, washed in PBS, then treated with 100ul of 0.1M glycine for 15 minutes at room temperature to block free aldehyde groups. The wells were then washed twice with assay buffer (PBS + 0.5% BSA), then left in 100ul of buffer for 30 minutes at room temperature. Patients' serum (50ul) was added at the appropriate dilution, and to this 50ul of antibody 3E1-2 was added. After 45 minutes incubation at 37°C, the wells were washed three times with assay buffer, then incubated with 10⁵ counts per minute/well of ¹²⁵I RAMG for 45 minutes at 37°C. Plates were washed four times with assay buffer, cut, and counted in a gamma counter. Results Details of Cell Fusion and Screening After cell fusion, hybrids were seeded into 384 microtiter wells and growth was observed in approximately 70%. The supernatants were screened by the rosetting assay on the lumphocytes and granulocytes of two normal individulas, to eliminate unwanted antibodies, then tested by the immunoperoxidase technique on formalin-fixed paraffin embedded sections of the immunising tumour. Several were selected for further characterisation on formalin fixed, paraffin embedded sections of normal breast and other breast tumours and one antibody, 3E1-2, is described herein. 3E1-2 was found to be of the IgM class and using the immunoperoxidase method on tissue sections had a titre on breast carcinoma cells of 1/2000 for supernatant and 1/20,000 for ascites. Reactivity of 3E1-2 with Breast Carcinoma The antigen detected by the 3E 1-2 antibody survived formalin-fixation and paraffin embedding, and so a large number of carcinomas of the breast, and normal tissues could be easily examined. Of the 37 carcinomas of the breast examined, all were positive for 3E1-2 (see Table 1). Various types of breast carcinoma were represented and included 23 ductal carcinomas, 4 adenocarcinomas, 4 comedocarcinomas, 1 lobular carcinoma and 5 metastatic carcinomas. Of the 37 specimens examined, 31 had between 60-100% of malignant cells staining, 5 had 10-60% of malignant cells staining and one showed less than 10% of malignant cells staining. Although each histological type stained, the distribution of staining varied. Ductal carcinomas showed staining of both membranes and cytoplasm, whereas adenocareinoma exhibited predominantly luminal membrane staining and little cytoplasmic staining. Comedo (in sutu) carcinoma showed staining of luminal membranes, as well as cytoplasmic staining, while lobular carcinoma stained predominatly cytoplasm, and metastatic carcinoma resembled the ductal carcinomas in staining pattern. In themajority of cases (>85%) the staining intensity of carcinoma of breast was either very strong or strong. Thus all carinomas were clearly positive, although the number of cells staining, and the distribution and intensity of staining varied with the individual tumours. Reaction with Normal, Hyperplasic, Lactating and Benign Breast Tissue To further characterise the tissue specificity of the 3E1-2 antibody, normal, benign, lactating and hyperplastic breast tissue was examined (Table 2). The immunoperoxidase technique revealed that 3E1-2 recognises an antigen present on normal breast tissue. However, it was restricted to only 60% of normal individuals tested, and on average, to 50% of ducts and acini in any one section. The staining was always of the liminal membrane surface, with the apical cytoplasm staining in some cases. This is in contrast to the pattern of staining seen in carcinoma of breast, where staining was cytoplasmic as well as membranous, with most sections showing greater than 80% of cells staining. Benign lesions of the breast such as papilloma, gynaecomastia and fibroadenoma showed almost exclusively staining at the luminal membrane surface of neoplastic epithelial cells, a pattern similar to that of normal breast. Cystic hyperplasia showed a staining pattern similar to the benign conditions. Lobular hyperplasia however exhibited not only luminal membrane staining but also a staining of the epithelial cell cytoplasm, and of the myoepithelial cells. Of interest was the finding that a section of prelactating hyperplasia showed no staining of any hyperplastic components, whereas adjacent normal ducts in the same sections showed intense luminal membrane staining. Sections of lactating breast exhibited variable staining, the pattern of staining of this condition is unclear. Thus normal breast tissue and benign conditions of the breast was also positive with the 3El-2 antibody, but the distribution of staining was quite different from that seen with breast cancer. Reaction with Normal Human Tissues Monoclonal antibody 3E1-2 showed a restricted reactivity with normal human tissues (Table 3). In the kidney strong membrane staining was observed in >95% of the distal convoluted tubule cells, but only 20% of those also gave strong cytoplasmic staining. Some collecting ducts exhibited luminal membrane staining, with weak cytoplasmic staining, whereas others did not react at all. Other significant reactions were only seen in the sweat glands of the skin, the epithelium of the urinary bladder and the subpopulation of lung alevolar cells. In the sweat glands the staining was restricted to the canaliculi between the cuboidal epithelial cells. The alveolar cells of the lung show membrane staining of the surface which faces the alveoli; the surface resting on the basement membrane did not stain. A moderate degree of staining was seen in the striated ducts of the salivary gland, the myocardium of the heart and the white matter of the brain and spinal cord. Weak staining (possibly background) was seen in the epithelium of the thyroid, stomach and ileum. Tissues showing no staining were liver, prostate, ovary, spleen, colon, adrenal gland, parathyroid gland, bone marrow, gall bladder, cervix pancreas, lymph node, ileum, stomach, trachea, salivary gland, heart, lung, kidney, skin and thyroid. Reactivity of 3E1-2 with other Human Tumours Monoclonal antibody 3E1-2 was tested on a ranqe of non-breast tumours to further determine the distribution of the "antiqen" detected by the 3E1-2 antibody (Table 4). Using the immunoperoxidase technique it was found that 3E1-2 recognises an antigen present on renal, endometrial and cervical carcinomas. Staining of malignant cells was both cytoplasmic and membranous and was of similar intensity to that seen on carcinoma of the breast. Lung carcinomas in the main appeared to be negative, but in some, a small proportion of cells in between the larger tumour masses were positive. It is probable that these are alveolar cells, as they constitute 10-30% of cells present, and are always present between large groups of non staining malignant cells. The fact that metastatic lung carcinoma shows no staining supports this belief. Examples of carcinoma of colon, malignant melanoma, nasopharyngeal carcinoma, basal cell carcinoma (skin), transitional carcinoma (urinary bladder), mucinous cystadenoma (ovary) and giant cell tumour of bone were found to be non-reactive with the 3E1-2 antibody. Reactivity with in vitro cell lines An extensive range of in vitro cell lines were examined with the 3 El-2 antibody (Table 5). The antibody failed to react with 10/10 breast carcinoma cell lines, with 5/5 colon carcinoma cell lines, or to any other in vitro cell line, including cells derived from melanoma, and tumours of the uterus, lung, pancreas, kidney, parathyroid, and fibroblastic or haemopoietic derived cell lines. As many of these cell lines were adherent, and trypsin was used for their preparation which conceivably could destroy a trypsin sensitive antigen, several of these cell lines were prepared by using EDTA; they still did not react with 3E1-2. In addition, histologic sections of breast carcinoma did not loose their ability to bind 3E1-2 when treated with varying concentrations of trypsin for up to 30 minutes at 37°C. Using the immunoperoxidase technique on air dried monolayers of in vitro cell lines have indicated that the 3E1-2 antibody reacts with cytoplasmic components of a few carcinoma of breast cell lines. Of those cell lines tested (T47D, BT-20, PMC-42, MCF-7 - derived from carcinoma of the breast, PMC-115, PMC-116 - fibroblastic origin), only two T47D and BT-20 reacted with the 3E1-2 antibody. Staining of T47D was restricted to the inner cell membrane of 5% of cells, while 95% of BT-20 cells gave strong cytoplasmic staining. Molecular Weight Estimation After cell surface labelling of fresh carcinoma of breast and normal kidney cells by both the lactoperoxidase and chloramine T methods, no specific bands were found on either 7.5% or 12.5% (reduced and unreduced) SDS-PAGE gels. Appropriate positive and negative control antibodies were used for all gels, as were molecular weight markers. In addition internal labelling of carcinoma of breast cell line BT-20 (which is membrane negative, cytoplasmic positive) with S³⁵ methionine produced no specific bands on either 7.5% or 12.5% non reduced SDS-PAGE gels. Serum Inhibition Study To determine whether 3E1-2 could be used for diagnosis, based on the detection of antigen in serum, a preliminary study was performed on twenty patients with carcinoma of breast. The serum from these patients was used to inhibit the binding of 3E1-2 to a membrane preparation. Serum from normal females (aged 35-65 years) showed a small degree of inhibition, implying that the "3E1-2 antigen" is normally present in serum, but only in small amounts (see Figs. 1 and 2 ) . By c o n tr as t, s e r u m f r o m t he pa t i e n ts w i th p r i m a r y carcinoma of breast and those with metastases showed significant inhibition of 3El-2 binding. Fig. 2 shows diagramatically an example of this inhibition, comparing it to serum from a normal female control (± standard deviation). These results were quantitated by taking the inhibition graph of each patient, and calculating the "inhibition units" (I.U.) per ml of serum (calculated as the reciprocal of the serum dilution giving 50% inhibition of the maximum binding of 3E1-2 to the membrane preparation x20). Serum from the patients with localised carcinoma of breast showed a significantly greater degree of inhibition, compared to aged matched female controls (p<0.005). This was also the case with patients having a primary tumour and metastases (p<0.01). Serum from patients with metastases inhibited almost twice as much as serum from patients with a localised lesion only. These two groups were also significantly different from each other (p<0.05). Discussion The production of the murine monoclonal antibody 3E1-2 and its extensive characterisation by the immunoperoxidase technique on normal and neoplastic tissue is described above. The antigen recognised by 3E1-2 survived formalin fixation and paraffin embedding, (a property selected for in the initial screening), and was found to be present on all carcinomas of the breast and most benign proliferations of the breast although in a different form. However only 60% of normal breast tissue reacted with the antibody, and the majority of normal human tissues were non-reactive. Using a surface membrane assay, no reaction was found with 43/43 in vitro cell lines, including 10/10 carcinoma of the breast cell lines. However 3E1-2 has subsequently been found to localise to the cytoplasm of two breast carcinoma derived cell lines. An important finding was that 3E1-2 detects an antigen present in serum, which is significantly raised in patients with carcinoma of the breast. In a new approach initial screening was performed on formalin fixed tissue, which led to the identification of an antigen which is resistant to formalin fixation and paraffin embedding. Indeed, this antigen also resists fixation by periodate-lysine-paraformaldehyde and glutaraldehyde, and as such, lends itself to conventional study by immunohistochemical means without having to resort to using fresh frozen tissue sections. In spite of this the nature of the epitope recognised by 3E1-2 is unknown, as is the molecule it resides on. Surface protein labelling experiments using both the 1 actoperoxidase and chloramine T methods, on known positive tissues, produced no specific bands by SDS-PAGE analysis. The explanation for this could be that the molecule is not protein and therefore will not label with ¹²⁵I, alternatively it could be glycoprotein but containing only a few exposed tyros ine residues, and therefore does not label sufficiently to be seen on the autoradiograph. To further study this internal labelling of the in vitro cell line BT-20 with S³⁵ methionine was performed, but again no specific bands on SDS-PAGE gels were obtained. It could be that methionine is present in limiting amounts, but it is not unlikely that it is glycolipid rather than protein particularly as cell membrane glycolipid molecules have been described elsewhere using a similar technical approach. Immunohistochemical studies have revealed that 3E1-2 recognises an antigen with a restricted distribution on normal human tissue. Indeed, 3E1-2 reacted predominantly with a subpopulation of epithelial cells. This phenomenon of antibodies recognizing epithelial "subsets" is not uncommon and has been previously reported for other monoclonal antibodies to human tumours. The antigen is present on the surface membrane of breast, kidney and lung epithelium, and also on the membrane and cytoplasm of carcinomas derived from these tissue types. All of these tissues are involved in specialised secretory functions and one could postulate a role for the 3E1-2 molecule identified in that process. Compared to other monoclonal antibodies produced to carcinoma of the breast, 3E1-2 reacts with very few normal human tissues. However, comparing antibodies is difficult, as few studies have fully characterised their monoclonal antibodies using sensitive immunohistochem ical techniques - and this is where the ability to test a collection of fomalin fixed tissues is so important. Interestingly, on normal breast tissue and on distal convoluted tubule of kidney, the "3E1-2 antigen" shows heterogeneity of staining, i.e. morphologically identical cells, located adjacent to each other, show considerable variation in staining pattern. This could indicate that the molecule identified may change its expression during the cell cycle, or in response to changing levels of hormones, chemicals or other substances required by the cell or indeed, that histologicalty identical cells have a different function. Such an hypothesis may explain the failure of 3E1-2 to react with surface components of in vitro cell lines which usually are grown in the absence of hormones and other substances normally present in vivo. Detailed examination of sections of breast, including normal, benign, hyperplastic and malignant tissues, have been informative in that each has shown the expression of the "3E1-2 antigen" during various stages of development, indicating that 3E1-2 defines an antigen important in identifying the different stages of development from normal to malignancy. For example, in normal breast, fibroadenoma, papilloma and gynaecomastia, the antigen was localised 'to the luminal membrane surface and apical cytoplasm of the epithelial cells. Hyperplastic conditions showed two types of staining. Cystic hyperplasia, a "benign" change of the breast limited to fibrosis, sclerosing adenosis and cyst formation, shows luminal membrane and apical cytoplasmic staining, similar to that of normal and benign breast conditions. Secondly, lobular hyperplasia which is essentially an epitheliosis, showed not only luminal membrane staining but also cytoplasmic staining, and stain'ing of myoepithel ial cells. This finding is relevant in that only in the case of lobular hyperplasia is there evidence for transition to malignancy and carcinoma formation. Few carcinomas develop from fi broadenoma, gynaecomastia and fibrocystic disease. Our results show clearly that 3E1-2 is localised to both the membrane and cytoplasm of malignant cells (Table 1). In this respect the 3Ε1-2 antibody confirms the link between lobular hyperplasia and carcinoma of the breast, and may be important in studying the genesis of carcinoma of the breast. 3E1-2 does not appear to recognise known components of human milk such as casien and lactalbumin, as immunoperoxidase staining of milk in histological sections was negative. The distribution of immunoperoxidase staining and failure of SDS-PAGE analysis of breast tissue discounts hormone receptors for estrogen, progesterone and prolactin as the target. In addition to this the antigen detected has a unique reaction on in vitro cell lines and tissue sections, when compared to a large range of carcinoma of the breast antigens described in the literature. There does however, appear to be some similarity between 3E1-2 and the LICR-LON M8 and F36/22 monoclonal antibodies but in our hands M8 and 3E1-2 have different tissue distributions and F36/22 reacts with the surface membrane of many other different cell lines. Thus 3E1-2 is a "new" breast carcinoma antigen previously undefined. Potentially the most important finding of this study is that the antibody 3E1-2 detects an antigen present in the serum of patients with carcinoma of the breast patients. The results show (see Figs. 1 and 2) that sera from patients with a localised tumour gave significantly more inhibition of 3E1-2, compared to normal age matched female controls. The same was true for patients with metastases. This compares favourably with all other carcinoma of the breast serum markers described to date. The value of the 3E1-2 marker can only be assessed when compared to other markers of breast cancer, used thus for such as CEA and the recently described anti-human MFGM. When this comparison is made, elevated levels of the 3E1-2 antigen occur in most (>85%) of patients with primary breast cancer and metastases (p<0.05). However, with CEA testing elevated levels occur in only 15% of patients with primary tumours and 65% of patients with metastases. For anti human MFGM these values are only 25% and 75% respectively. Although this work is at a preliminary stage the results were significant (p<0.05). The current test, a radioi mmunoassay, can be performed rapidly on large numbers of serum samples and is highly reproducible. In an alternative embodiment of the invention, the 3E1- 2 antibody is used to localise carcinoma of the breast in vivo. This embodiment will now be described in greater detail. Methods Antibodies

The monoclonal anti-breast antibody used was 3E1-2 (IgM), which reacts strongly with membrane and cytoplasm of breast carcinomas and with the luminal membrane of normal breast. Control monoclonal antibodies to murine antigens Thy-1.2 (IgM) and Ly-1.1 (IgG 11) were also used, and in several cases, the patients own IgG was isolated using protein A affinity chrom atography; an anti-colon carcinoma monoclonal antibody was also used: none of these control reagents reacted by the immunoperoxidase method with normal or malignant breast tissue. The IgM antibodies were isolated from ascites fluid by dialysis against distilled water at 4°C after which the precipitate was collected; IgG antibodies by NH₄(SO₄)₂ precipitation. Antibody activity measured either by the rosetting technique referred to above, or by the immunoperoxidase method on sections of breast carcinoma. The purified antibodies were retested for activity after all procedures, filtered through a 0.22um filter (Gelman Acrodisc, Michigan, USA), and tested for pyrogens,. and sterility prior to use. Radio Iodination

Purified monoclonal antibodies (100g) were labelled with ¹³¹I using chloramine T, and the ¹³¹I labelled antibody separated from free iodine on a PD-10 column (Pharmacia Chemicals - Sweden). Control antibodies or human IgG were labelled with ¹²⁵I by the same method. Patients

All patients (except for one with lymphoma) had proven breast carcinoma, and their clinical details are shown in Table 6. One hour before the subcutaneous (sc.) injection of monoclonal antibody each patient received potassium iodide (5 ml at 16.2% w/v), sodium perchlorate (400mg) orally and an intravenous injection of Hydrocortisone (100mg) and promethazine (25mg). Radio Imaging

Approximately 24 hours after the injection of the radiolabelled antibodies the patients were scanned using a Toshiba GC 402 A gamma camera and a high energy parallel hole collimator; computerized acquisition using an Informatek Simis 4 computer was performed. A window setttng of 50 KeV with a 50% window was used for ¹²⁵I S c a n s and 360 KeV with 20% window for ¹³¹I scans. There was 15% down scatter of ¹³¹I photons into the ¹²⁵I channel but since ¹²⁵I counts were three times greater than those from ¹³¹I, this amounted to only 10% of counts. Blood pool subtraction was not performed because of the superficial site of the gland. In three patients comparison of the uptake of specific antibody and non-specific immunoglobulin control in the ly mp h no d e s w a s pe r f o r m e d by u s i ng ^{1 3 1} I -3 E 1 -2 ( s pe c i f i c antibody) and as control ¹²⁵I-Thy-1.2 (IgM); ¹²⁵I-Ly-1.1 (IgG), ¹²⁵I-30.6 and ¹²⁵I-homol ogous IgG. The fraction of radiolabelled antibody localising in the axillary nodes (F) was estimated by measuring the nodal uptake (N) using the alpha camera, corrected for "down scatter" from the ¹³¹I channel and comparing this to the amount of radiolabeoled antibody found at the time of the imaging (Im) and time of injection (In) so that F = N xA, where A = the correction factor (2.5) used for ¹²⁵I attentuation in the axilla, determined by placing a source bottle under the patients armpit and comparing this to the actual radioactivity in the bottle (ie 400 vs 1000 cpm). Results

In this embodiment patients were injected with radiolabel led 3E1.2 antibody by the subcutaneous route into the digital web space. The rationale of this study was that the antibody would enter the lymphatics and pass through draining lymph nodes and detect metastatic tumour residing there. The results are shown in summary in Table 6. Discussion

The currently available techniques for the detection of breast cancer include palpation, x-ray mammography, thermography, ultrasound, diaphranography and nuclear magnetic resonance. However, these methods distinguish between benign and malignant lesions with difficulty. It is considered to be important to be able to detect metastases in draining axillary lymph nodes in carcinoma of the breast as this serves to distinguish the early and late stages of the disease and currently is used to design appropriate therapy. With this in mind we have attempted to detect small metastases in the draining axillary lymph nodes of patients with breast cancer. The method of using radiolabelled monoclonal antibodies continued with scintography has recently been receiving some attention for the detection of a variety of tumours. The re-emergence of this technique - attempted unsuccessfully some years ago - has been developed because of the advent of monoclonal antibodies with at least some degree of specificity for tumours. Thus using the intraveneous route we and others have been able to detect both primary and secondary tumour arising from a range of different tumours. At this time it is not clear whether very small tumours can be detected and there is clearly a need for the refinement in the use and production of these monoclonal antibodies, the type of radiolabelling and the site and route of injection. The antibody used herein, 3E1.2 was used with little success by the intraveneous route to delineate secondary deposits from breast cancer. The results obtained (data not shown) were similar to those described using other antibodies. However it appeared that small tumours would evade detection by intravenous means - the major problem being the high content of circulating radiolabelled antibody which, after subtraction, leaves little detected in the tumour. To overcome these problems and to determine whether subcutaneous antibody administration was a better route to the draining lymph node via the lymphatics, we used radiolabelled antibody given subcutaneously. In the patients studied we were clearly able to detect metastases and to confirm their presence in those patients who, by other means were already suspected to have tumour involving axillary lymph nodes. Thus the method demonstrates that radiolabelled antibody can react with tumour in draining lymph nodes. Of greater importance however was the finding of additional lymph nodes which had not been clinically palpable and those whose presence was not suspected i.e. the immunoscintography test was more sensitive than conventional clinical examination for the detection of metastases in draining nodes. Whether this technique is more sensitive than conventional lymphangiography remains to be seen. None-the-less, this is one of the first examples where monoclonal antibodies appear to be superior to the existing commonly used clinical method for the detection of small tumour deposits. Several comments should be made regarding the method. Firstly, it did not appear necessary to use subtraction with a non-specific antibody as the specifically labelled antibody was retained in the abnormal lymph node but rapidly cleared from the normal lymphatics and normal nodes. Secondly, the kinetics of the disappearance of antibody were such that scanning between 16-24 hours appeared to be more suitable than earlier or later scans. Thirdly, it was clear thai: "pathological" lymph nodes retained radiolabelled antibody both specifically (when involved with carcinoma of the breast) and non-specifically (when involved with other diseases such as lymphoma). This is not unexpected because of the greater vascularity of involved lymph nodes and because of the possibility of some hold up in lymphatics draining in these nodes. In this context, is the immunoscintography described herein any different than lymphangiography using more conventional reagents. We feel the method does have advantages over the conventional lymphaπgiogram in that normal lymph nodes are not detected; there is twice as much (or more) specific antibody detected, compared with non-specific and in a grossly abnormal lymph node the collection of antibody was not greater than that found in other tissues. However the sensitivity of the method needs improvement, for example rad io iodi nated antibody is not the best method for labelling because of the short half life in vivo of such antibodies. However it should be noted that in recent studies in both mice and guinea pig, tumours of 2.3mg could be detected and there is no reason why such a sensitivity should not be obtained in the type of studies described herein. However, whether the method will approach that to detect micro-metastases in bone marrow will depend on the amount of antibody given, the quality and quantity of the radiolabel, and the detection system. Possibly the combined use of several antibodies to different epitopes on the cell surface may lead to a magnification and greater sensitivity. However, the essential question to ask is whether this procedure - both simple and painless to perform - represents a real advance in the early diagnosis of breast cancer. We believe an advance has been made since no other technique is currently available which can so readily distinguish between normal and malignant nodes. We have therefore shown that (i) palpable nodes can be detected by immunoscintography, (ii) normal nodes do not take up the antibody, (iii) abnormal nodes (lymphoma) only marginally retain antibody, (iv) localisation of the antibody is specific by at least a factor of two, (v) nodes not clinically palpable can be detected, (vi) the subcutaneous site of injection is superior to the intravenous route for this antibody. In the few patients studied, palpable lymph nodes were readily detected but impalpable nodes were also found. What is now required is a larger prospective study where patients, considered clinically to be stage 1 (carcinoma localised to breast only) are subjected to immunoscintography prior to operation, and the immunoscintography findings compared with subsequent histological-exam ination of axillary lymph nodes. At this time there are a number of patients who clinically appear to be stage 1, but at operation are found to be stage 2 or stage 3. If this technique alters the management of these patients and reduces the extent or necessity for mastectomy, combined with an earlier diagnosis, the technique can then be pronounced to be of use.

Claims

1. A cell line w h ic h produces a monoclonal antibody ha vi n g a predominant s p e c i f ic it y to breast cancer antigens.

2. A continuous cell line which produces breast cancer antibodies identified as clone 3E1-2 which produces breas carcinoma antibodies in vitro in hypoxanthine-aminopterinthymidine medium comprising a fused cell hybrid of human adenocarcinoma-primed BALB/c mouse spleen cells, and NS-1 mouse myeloma cells, said antibodies reacting with a breast tissue antigen.

3. A monoclonal antibody having a predominant specificity to breast cancer.

4. A method of detecting breast cancer antigens in blood serum secretions or tissue comprising the step of applying to a blood serum or secretion to be tested a monoclonal antibody having a predominant specificity to breast cancer.

5. The method of Claim 4, wherein said monoclonal antibody is applied, using the immunoperoxidase technique, to formalin-fixed or fresh neoplastic tissue sections, and detecting the presence or absence of showing uniform staining of the cytoplasm and membrane.

6. The method of Claim 4, wherein said monoclonal antibody is applied, using a radioimmunoassay technique, to blood serum, and detecting the presence or absence of significant inhibition of monoclonal antibody binding to a membrane preparation.

7. The method of Claim 4, wherein said monoclonal antibody is applied, using an ELISA technique, to blood serum, and detecting the presence or absence of significant inhibition of monoclonal antibody binding to a membrane preparation.

8. The method of Claim 4, wherein said monoclonal antibody is radiolabelled and injected intravenously and detecting the presence or absence of axillary lymph nodes by immunoscintography.