Pigmentation Associated, Differentiation Antigen of Human Melanoma and Autologous Antibody

The present invention concerns an antigen associated with pigmented melanoma cells and the autologous antibody which recognizes it.

Sera from tumor patients may contain antibodies reacting with their own tumor cells which may be detected using suitable sensitive serological assays such as the mixed hemadsorption assay (MHA). When the cells and the serum come from the same individual this is known as autologous reactivity; when the serum is tested on cells from a different individual then the reaction is known as an allogeneic reactivity. The analysis of the specificity of antibodies in an autologous reaction is easier than in an allogeneic reaction because reactions with unrelated allogeneic antibodies are eliminated. This approach to study tumor cells (autologous typing) has led to the detection of a number of antigens of melanoma and other tumors that have very restricted distributions.

One approach in the search for human tumor-associated antigens has been to test sera from tumor patients for reactivity against the autologous tumor cell line. A possible disadvantage of this approach is that antigens which elicit cellular immunity but not antibodies will not be detected. This point is significant since in animal models tumor-specific transplantation antigens, which elicit strong cell-mediated immunity, do not in general induce antibodies. Still if antibodies reacting with autologous tumor cells are found, they provide a reagent that can be tested for specificity with a thoroughness not readily attainable using assays for T cell immunity. Several interesting antibodies have been defined by this approach. Some of these unique antigens are found only on an individual tumor; they have been designated Class 1 antigens. A detailed biochemical characterization of these antigens has been difficult due to the fact that none could be detected by immunoprecipitation. Thus antibodies to the Class 1 antigen AU have been found in serum of patient AU (Carey, et al. Proc. Nat'l. Acad. Sci., U.S.A., 73:3278 (1976)), but characterization of the antigen has been difficult since the antibodies are non-precipitating. Accordingly, efforts have been exerted to locate precipitating autologous antibodies to tumors.

The present invention concerns an autologous precipitating antibody and the gp70 pigment-associated antigen on melanoma cells which it recognizes. The antibody is useful in detecting pigmented melanoma cells in excised specimen, serum or urine.

Adherent cells from human tumor cell lines were maintained in Eagel's Minimum Essential Medium supplemented with 2.5% fetal calf serum, 5% newborn calf serum, 100 units/ml penicillin and 1 mg/ml streptomycin. For passage, cells were detached with 0.1% trypsin, 0.02% ethylene diamine tetra-acetic acid in Hank's salt solution without divalent cations. Non-adherent cells were cultured in RPMI 1640 medium supplemented similarly except with 7.5% fetal calf serum. Cultures were regularly tested for mycoplasma and contaminated cultures discarded.

Monolayers of cells from three 150 cm.sup.2 flasks were washed 4 times, then scraped off and pelleted. Cells were solubilized in 5 ml 0.15M NaCl, 0.01M Tris HCl pH 7.2,) 0.01M MgCl.sub.2, 0.5%NP40 detergent, 2 mM phenyl-methylsulfonyl fluoride (PMSF), 20 units/ml aprotinin. After passing repeatedly through a 22 gauge needle, followed by a 26 gauge needle, the mixture was incubated for 15 minutes at room temperature, then 15 minutes at 4.degree. C., centrifuged for 30 minutes at 100,000 g and the supernatant collected. This material was fractionated on a 1 ml Con A-Sepharose column, using 0.15M NaCl, 0.01M Tris HCl pH 7.3, 0.1% NP40 as column buffer, and eluting with 0.2M .alpha.-methyl D-mannoside. Samples of 20 .mu.g were iodinated with 1.0-2.5 mCi with chloramine. Protein-conjugated .sup.125 I was estimated by precipitation with cold 10% (wgt/vol) trichloroacetic acid followed by washes with ethanol and acetone.

For metabolic labeling, a confluent 150 cm.sup.2 flask was incubated with 0.5 mCi [.sup.35 S]methionine in 10 ml methionine-free Minimal Essential Medium containing 10% heat-inactivated fetal calf serum that had been dialyzed against Dulbecco's PBS. A solubilized preparation was obtained as described above.

Aliquots of 5.times.10.sup.5 protein .sup.125 I cpm were used routinely. 0.2 ml of medium was added to each sample, then 3 l or 15 l of serum. After 18 hours at 4.degree. C., 15 l of washed, packed Staph A was added in 1.0 ml and incubated 30 minutes at 4.degree. C. After centrifugation (5 minutes at 3000 rpm), the pellet was washed four times. Pellets were assayed for radioactivity, then extracted with 45 l 0.01M Tris HCl pH 7.2, 2% SDS, 12.0 mg/ml dithiothreitol, 15% (wgt/vol) sucrose, 0.01% pyronin Y by heating 5 minutes at 100.degree. C. After centrifugation, supernatants were used in acrylamide gel electrophoresis using 9% slab gels. Dried gels were exposed to Kodak XAR film with a DuPont Lighting Plus Intensifying Screen for 2-25 days. Molecular weight standards were myosin (212,000), .beta.-galactosidase (116,000), phosphorylase (97,500), bovine albumin (66,000), ovalbumin (43,000), Concanavalin A (27,000) and myoglobin (17,000). The standards were labeled, with formaldehyde and tritiated borohydride (5-8 Ci/mole). For 2-dimensional electrophoresis (isoelectric focusing followed by SDS-electrophoresis) immune precipitates were extracted. For unreduced samples, DTT was omitted and 14.0 mg/ml iodoacetamine was added to samples.

To lectin affinity beads, samples were applied and, after running through 10 column volumes, were eluted with the appropriate sugar at 1.0M.

Antibodies were found in serum from a melanoma patient, AU, that precipitate an antigen, of molecular weight 70,000, form an .sup.125 I-labeled NP40 solubilized preparation from the autologous tumor cell line SK-MEL-28. These precipitating antibodies were absorbed by SK-MEL-28 cells but not by cells from another melanoma, SK-MEL-37. Additional absorptions were performed to determine the distribution of this antigen. Eight of 15 melanomas absorbed and antibodies and 17 other cell types were negative, including 3 astrocytomas, 3 ovarian carcinomas, 1 uterine carcinoma, 2 bladder carcinomas, 1 renal carcinoma, 2 lung carcinomas, 1 breast carcinoma, 1 colon carcinoma, 1 choriocarcinoma, 1 leukemia and human erythrocytes.

A positive absorption among melanoma cell lines was associated, without exception, with pigmentation of the cell lines. To further examine this relationship, quantitative absorptions were performed with 2 cell lines: SK-MEL-28, a lightly pigmented cell lines, and SK-MEL-23, a very highly pigmented cell line. Results demonstrate that SK-MEL-23 is approximately 30 times more effective than SK-MEL-23 at absorbing these antibodies. This experiment was repeated with a second very highly pigmented melanoma (SK-MEL-21), with similar results, with 2 astrocytomas showed negative absorption. Therefore, and the correlation of antigen expression with pigmentation was confirmed. This antigen was therefore designated pigmentation associated antigen (PAA). PAA is distinct from melanin itself which is not solubilized under the conditions employed.

Iodinated antigen preparations from two other lines, the highly pigmented SK-MEL-23 and the non-pigmented SK-MEL 37 were also observed. AU serum precipitated nothing from labeled SK-MEL-37 but precipitated a component corresponding to PAA from labelled SK-MEL-23, and the band seen was much stronger than that seen with labeled SK-MEL-28. This experiment further supports the correlation of PAA expression with degree of pigmentation of the cell line.

Also, PAA was precipitated from iodinated antigen preparations from normal human melanocytes, indicating that it is a normal differentation marker.

SK-MEL-28 melanoma cells labeled by [.sup.35 S]methionine or [.sup.3 H]glucosamine had previously been examined for the presence of an antigen that precipitates with autologous serum, and no precipitated antigen was detected. Since .sup.125 I labeling, in contrast to biosynthetic labeling methods, will label components of the fetal calf serum included in tissue culture medium, it is important to exclude the possibility that the antigen is a heterologous component derived from the serum. Immune precipitation was not inhibited by a large excess of unlabeled fetal calf serum. Also, a Con A eluate fraction of fetal calf serum was labeled with .sup.125 I and tested in immune precipitation with serum AU: no component resembling PAA was precipitated. Therefore, we conclude that PAA probably is a biosynthetic product of the tumor cells.

Other characteristics of PAA have been determined. Unreduced samples migrated at the same rate as reduced samples on SDS-acrylamide gels. The isoelectric point is at pI 5.2, with a second minor spot at pI 5.5. The antigen binds to Concanavalin A-Sepharose and is eluted with methyl-a-D-mannoside. A fraction (10-25%) also binds to wheat germ agglutinin and is eluted with N-acetyl-glucosamine. A similar fraction binds to Ricinus communis lectin-I and is eluted with D-galactose. PAA did not bind to the other lectins tested, (lentil lectin, peanut agglutinin, soybean agglutinin, and lectins from Bandeira simplicifolia, Helix pomatia and Lotus tetragonolobus).

The frequency of antibodies to PAA was investigated by screening 96 sera samples from melanoma patients using the same precipitating method with .sup.125 I-labeled SK-MEL-28. Fifteen .mu.l of each serum was tested, while 3 l of AU serum has been used routinely. None of the sera precipitated PAA. A smaller number of normal human sera were also tested and found to be negative. Precipitating antibodies were found in approximately 10% of these sera, but the antigens recognized were diverse and different from the 70K antigen.

It is critical in these experiments that the antigen preparation be in .sup.125 I-labeled eluate fraction from a Concanavalin A-Sepharose column of an NP-40-solubilized whole cell extract. When the same experiments were performed with an iodinated solubilized crude membrane preparation, no precipitation was seen. This emphasizes the fact that variation in sample composition can be critical in determining whether or not a particular antigen is detected. Moreover, no specific components were precipitated by the serum from lysates of cells metabolically labeled with [.sup.35 S]methionine a [.sup.3 H]glucosamine. Sensitivity for detecting a particular antigen would seem to depend on the efficiency of labeling that particular molecule relative to other cell constituents, to the amount of radioactivity used for each sample in immune precipitation, to the background seen on acrylamide gels due to nonspecific sticking to the immunoabsorbent, Staphylococcus aureus, and to the film exposure time. The antigen identified is a differentiation antigen closely correlated with pigmentation, highly pigmented melanomas express more PAA than do lightly pigmented melanomas. PAA was undetectable in non-pigmented melanomas and other tumor cell types.

Other melanocyte differentiation markers share some properties with PAA, but seem to be distinct. Tyrosinase, the major enzyme involved in melanin production, is a glycoprotein of approximately 75,000 dalton. But an antiserum that reacts with human tyrosinase, did not precipitate PAA.

Antibodies to other melanocyte differentiation antigens have been found in human sera. The Mel 1 antigen, detected on normal melanocytes and fetal fibroblasts but absent from adult fibroblasts, normal kidney epithelial cells, and EB virus-transformed B cells, was recognized by IgG antibodies in 1 serum sample of 106 normal male sera tested. The AH antigen is present on most melanomas and astrocytomas but absent from carcinomas, fibroblasts and EB-virus-transformed B cell lines. These two antigens are clearly distinct from PAA, but the mechanism of induction of such antibodies may be similar.

Antibodies in the serum of melanoma patient AU precipitate an antigen from .sup.125 I-labeled extracts of cultured autologous melanoma cells. The antigen is present in other pigmented melanomas but not in non-pigmented melanomas or other tumor cell types, and the amount of antigen is correlated with the degree of pigmentation. These conclusions are based on absorption experiments with 11 pigmented melanomas, 8 non-pigmented melanomas, 3 astrocytomas, 12 carcinomas of various histological types, 1 leukemia and human erythrocytes. The antigen has a molecular weight of 70,000 and isoelectric point of pH 5.3. It binds to Concanavalin A-Sepharose. Ninety-six other melanoma sera were examined and none of them precipitated this antigen. The serum from patient AU also has antibodies to a unique (Class 1) tumor antigen found only on AU melanoma cells. The pigmentation-associated, differentiation antigen and the unique autologous antigen are clearly different in their distribution, but some relationship between these unusual antibody responses is possible.