Oct 10, 1986 - for residual antibody reactivity on SK-MEL-23 target cells. Glycolipids. ..... Thompson, K. M., Melamed, K., Eagle, B. D., Gorick, T., Gibson, T.,.
Proc. Nati. Acad. Sci. USA Vol. 84, pp. 2416-2420, April 1987 Immunology
Cell-surface antigens of melanoma recognized by human monoclonal antibodies (gangliosides/Epstein-Barr virus transformation/cancer immunology)
HIROSHI YAMAGUCHI, KOICHI FURUKAWA, SHEILA R. FORTUNATO, PHILIP 0. LIVINGSTON, KENNETH 0. LLOYD, HERBERT F. OETTGEN, AND LLOYD J. OLD Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021
Contributed by Lloyd J. Old, October 10, 1986
hybridization with mouse or human partners in the capture of human lymphocytes producing melanoma-reactive antibody. By combining the two techniques-i.e., EBV transformation followed by fusion with mouse NS-1 myeloma-we have been successful in generating a panel of human mAbs reactive with cell-surface antigens of melanoma.
ABSTRACT Human monoclonal antibodies (mAbs) were derived from lymph node lymphocytes and peripheral blood lymphocytes (PBL) from patients with melanoma. Four methods for generating human mAbs were compared: fusion with human [LICR-LON-HMy-2 (LICR-2)] or mouse (NS-1) cells; transformation by Epstein-Barr virus (EBV); and EBV transformation followed by NS-1 fusion. NS-1 fusion with lymph node lymphocytes resulted in a higher number of growing hybrids than LICR-2 fusion. Virtually no hybrids were obtained from NS-1 or LICR-2 fusions with PBL. EBV transformed lymphocytes from lymph node and peripheral blood with equal efficiency, and the yield of proliferating cultures for antibody screening was more than 10- to 30-fold greater than that obtained by fusion techniques. However, once antibodyproducing cultures had been identified, stability and clonability of EBV-transformed cells were poorer than that of NS-1 hybrid cells. To combine the strengths of both methods, cultures of EBV-transformed cells were fused with NS-1, and hybrid clones were isolated that showed vigorous growth, clonability, and stable antibody secretion. Detailed specificity analysis of the mAbs produced by six of these clones indicated detection of a class 1 (unique) melanoma antigen, a class 3 melanoma antigen, and four ganglioside antigens (GD3, GM3, and two other, as yet uncharacterized, heterophile antigens).
MATERIALS AND METHODS Tissue Culture. Tumor cell lines were established and maintained as described (10). Several melanoma cell lines (see Table 1) were adapted to growth in serum-free insulin (5 ,ug/ml)/transferrin (5 ,ug/ml)/selenium (5 Ag/ml) (ITS) medi-
um. Fusion Procedure. Lymph node lymphocytes (LNL) and peripheral blood lymphocytes (PBL) were isolated by Ficoll/ Hypaque centrifugation and fused with human LICR-LONHMy-2 (LICR-2) lymphoblastoid cells or mouse NS-1 myeloma cells as described (19, 25). NS-1 fusions with EBVtransformed cells (after two limiting dilution passages of 1-104 cells per passage) or LICR-2 hybrids were performed in the same way as direct lymphocyte fusions, but hybrids were selected by 0.2 mM hypoxanthine/0.4 ,uM aminopterin/ 32 ,4M thymidine (HAT) and 10 ,M ouabain (Sigma) for the first week and HAT for the second week; this was followed thereafter by culture in RPMI 1640 medium containing 15% fetal bovine serum, 0.2 mM hypoxanthine, and 32 utM
Several groups have reported specific humoral and cellular immune reactions to melanoma cell-surface antigens (1-9). In our laboratory, the reactivity of serum with surface antigens of cultured melanoma cells from the same patient (termed autologous typing) has been analyzed in >200 patients. Three classes of antigens have been defined in this way (2). Class 1 (unique) melanoma antigens are restricted to the autologous melanoma; six examples of class 1 antigens have been detected (10-15). Class 2 melanoma antigens are detected on the autologous melanoma, on a subset of allogeneic melanoma cells, and on other neuroectodermally derived tumors, and one of the best-analyzed class 2 melanoma antigens is the ganglioside GD2 (16). Class 3 melanoma antigens are not restricted to any differentiation lineage and are more widely distributed. The advent of methods for producing human monoclonal antibodies (mAbs) through immortalizing human lymphocytes with Epstein-Barr virus (EBV) (17) or by fusion with human or mouse lymphoblastoid/myeloma partners (18) provides a new level of precision in the analysis of the immune response to melanoma. Houghton et al. (19), Irie and co-workers (20-22), and other groups (23, 24) have isolated antibodies from melanoma patients reacting with melanoma cell surfaces or intracellular antigens. In the present study, we compared the efficiency of EBV immortalization and
thymidine. EBV Transformation. EBV-containing supernatants from the B95-8 marmoset lymphoblastoid cell line (26) were collected and passed through a 0.45-,um filter and stored at -800C. LNL or PBL were suspended in RPMI 1640 medium containing 10o fetal bovine serum and adjusted to 2 x 106 cells per ml. Five volumes of lymphocyte cell suspension was mixed with 1 vol of B95-8 supernatant and placed in T-30 flasks (Falcon). After incubation overnight in a 5% CO2 (in air) incubator, cells were washed once, resuspended at 1-4 X 105 cells per ml, and plated (2-8 x 104 cells per well) in 96-well tissue culture plates. Wells containing antibody-producing cells were expanded in 24-well tissue culture plates and subsequently subcultured by limiting dilution in 96-well plates. Serological Assays. Techniques for the detection of cellsurface antigens (11, 14) and intracellular antigens (19, 25) have been described. Antigen characterization by heat treatment and neuraminidase treatment was performed as described (27, 28). Antibody inhibition tests were carried out by mixing the cell extract or purified gangliosides with culture supernatant (diluted three doubling dilutions below the end Abbreviations: mAb, monoclonal antibody; LICR-2, LICR-LONHMy-2; EBV, Epstein-Barr virus; ITS, insulin/transferrin/selenium; LNL, lymph node lymphocytes; PBL, peripheral blood lymphocytes; HAT, hypoxanthine/aminopterin/thymidine; IA, immune adherence; PA, protein A; IF, immunofluorescence.
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 2416
Immunology: Yamaguchi et al.
Proc. Natl. Acad. Sci. USA 84 (1987)
plating of EBV-transformed cells resulted in a loss >80% of antibody-secreting cultures. However, 17 stable cultures of EBV-transformed cells secreting cell-surface-reactive antibody were derived. To overcome problems of instability and low cloning efficiency of EBV-transformed cultures, we attempted to develop stable NS-1 hybrids of antibodysecreting EBV-transformed cells. The fusion frequency was generally low (5.0 per 107 EBV-transformed cells), but hybrids secreting antibody with the same reactivity as the parental EBV-transformed cultures were obtained in 15/35 attempts. With vigorous subcloning, 8/15 clones retained stable antibody-producing capacity, a frequency similar to our experience in deriving stable NS-1 hybrids in this and past studies (25). Specificity Analysis ofHuman mAbs. Six human mAbs were chosen for detailed specificity analysis. The pattern of cell-surface reactivity in direct tests and absorption analysis with a panel of cultured cells is shown in Table 2. Five of the antibodies were produced by NS-1 hybrids with EBVtransformed cells; the other antibody (2.39M) came from an NS-1-LICR-2 cloned hybrid. GXMI. GXM1 is an IgM antibody derived from PBL of patient GX that reacts with the patient's own melanoma cell line, SK-MEL-177 (titer, 1/2048). With the exception of autologous melanoma reactivity and low titer reactivity with 1 of the 13 other melanoma cell lines, no other cell type, including autologous PBL or autologous EBV-transformed B cells, reacted in direct tests or absorption analysis. The antigen is heat labile but has not been further characterized. HJMJ. HJM1 is an IgM antibody derived from PBL of patient HJ. In direct tests this antibody reacts with 13/14 melanoma cell lines but not with 50 other cell types. Qualitative absorption tests with cells collected by mechanical scraping showed a positive absorption pattern with all cells except erythrocytes. Indirect immunofluorescence (IF) tests with fixed cells to identify cytoplasmic antigens indicated that all nucleated human cells were antigen-positive. To test the possibility that antigen-absorbing cell-surface reactivity was derived from intracellular antigen released by cell damage during collection, we compared the absorbing capacity of cells harvested by scraping (cell viability, 95%) (Fig. 1). In contrast to the results with cells harvested mechanically, absorption tests with trypsinized cells correlated with surface reactivity-i.e., antibody was absorbed with cells having surface expression of the antigen and not absorbed with cells lacking surface expression. The antigen detected by HJM1 is heat-stable and neuraminidase-sensitive, suggesting that it is
point), incubating for 30 min at room temperature, and testing for residual antibody reactivity on SK-MEL-23 target cells. Glycolipids. Cells were extracted by chloroform/methanol, and neutral and acidic glycolipids were isolated as described (28). GM3 and GD3 were purified in this laboratory. GM2 was prepared as described (29). GM1, GD1a, and GT1 were purchased from Supelco (Bellefonte, PA). GD2 was a generous gift of H. Wiegandt (University Marburg, F.R.G.). GD1b was kindly given by R. K. Yu (Yale University, New Haven, CT). ELISA for glycolipids and immunostaining after TLC were performed as described (28-30).
RESULTS Generation of Human mAbs. Four approaches to generating human mAbs from the lymphocytes of patients with melanoma were compared (Table 1). Fusion of LNL with mouse NS-1 myeloma cells resulted in a higher yield ofclones than fusion with LICR-2. NS-1 hybrids grew more vigorously than LICR-2 hybrids, and this facilitated clonal selection and expansion of NS-1-derived clones. No LICR-2 hybrids were obtained after fusion with PBL, and the frequency of NS-1 hybrids was extremely low with lymphocytes from this source. In contrast to fusion techniques, EBV transformation resulted in an equally high frequency of proliferating cells from LNL and PBL. Testing individual wells for immunoglobulin secretion after initial plating of fused or EBVtransformed lymphocytes indicated that NS-1 or LICR-2 hybrids were clonally derived, producing a single heavy chain class, whereas EBV transformation resulted in polyclonal expansion of B cells with >95% wells containing cells secreting IgM, IgG, and IgA. To identify cells producing antibodies with cell-surface reactivity, supernatants containing .500 ng of immunoglobulin per ml were allowed to react with a screening panel of 20 different cell lines, including 10 melanomas, 5 leukemias, and 5 epithelial cancers. In 22 cases, autologous melanoma cells from the lymphocyte donor were also available for screening. As shown in Table 1, supernatants from 1.1% of wells (10/895) from LICR-2 fusions with LNL contained surfacereactive antibodies; in the case of NS-1 fusions, the figure was 1% (5/509). Three stable antibody-secreting clones were derived from LICR-2 fusions and three were derived from NS-1 fusions. Primary screening of supernatant from EBVtransformed cells identified 2-3% of wells with surfacereactive antibody against allogeneic cells and 0.3-0.4% against autologous cells. Expansion and limiting dilution
Table 1. Generation of human mAbs from lymphocytes of melanoma patients Wells with growing cells, no. per 107 lymphocytes % IgC wells* Trials, Method/ Median Median no. (Range) (Range) lymphocyte source LICR-2 fusion 76 11.2 LNL 20 (21-94) (1.1-200.0) 10 0 PBL (0) NS-1 fusion 19 25.2 12 LNL (3-32) (6.3-37.3) 0 0 PBL 18 (0-8) (0-27.6) EBV transformation 100 270.0 LNL 13 (56-100) (31.0-500.0) PBL
38
300.0
(33.3-480.0) (0-76.7)
2417
100
(85-100)
Wells screened
Positive
for cell-surface
wells,
Stable cultures
reactivity,t no.
no.
or clones, no.
895
10
3
Allo, Allo,
509 21
5 0
3
f Allo,
613
I Allo,
1,899
18 8 38 46 258
6 0 5 6 8
Allo,
1Auto, 1,817
Auto, 15,288 1,505
5.0 35 NS-1 fusiont *% of wells with growing cells having Ig in culture supernatants of .500 ng/ml. tAllo, reactivity with allogenic target cells. Auto, reactivity with autologous melanoma cells. tFusion with EBV-transformed B cells or LICR-2 clones producing antibody to cell-surface antigens.
2418
Proc. Natl. Acad. Sci. USA 84
Immunology: Yamaguchi et al.
derived from lymphocytes of melanoma patients GXM1 HJM1 32-27M T A T A Cells T A Renal cancer 256 + SK-RC-6 + 512 + SK-RC-7 + 256 + SK-RC-9 -_ SK-RC-54 128 + SK-RC-45, -48 1024 + Bladder cancer - - - + 32 + T-24 32 + 235-J 512 + 5637, Scaber -_ Lung cancer 512 + SK-LC-8 - + SK-LC-6, -12 16 SK-LC-7 1024 + Breast cancer 128 MCF-7 + MDA-MB-361 64 MDA-MB-231 512 + CAMA SK-BR-5 + 128 + Colon cancer + 256 + - + HT-29, SW-480 + SW-403, SW-837 + 128 + Other cancers + CAPAN-2 ME-180 + SK-OV-3 + 256 + GCC-SV PBL (n = 4)
Table 2. Serological characterization of six human mAbs GXM1 HJM1 FCM1 DSM1 Cells T A T A T A T A Melanoma - - 64 + 512 + 10,000 + SK-MEL-13 - -1024 + 1024 + SK-MEL-23 - - 64 + 128 + 40,000 + SK-MEL-28 64 256 40,000 + SK-MEL-28* 16 + 32 + 32 + 40,000 + SK-MEL-29 4+ 2,560 + - + + SK-MEL-31 - + 4 + 64 + SK-MEL-37 - - 128 + 256 + 2,560 + SK-MEL-61 - - ± + 128 + 640 + SK-MEL-93-II 640 + SK-MEL-93-II* - - 256 + 4096 + SK-MEL-94 T21 - - 32 + + 10,000 + SK-MEL-130 - - 64 + 256 + 10,000 + SK-MEL-173 E.048+ - + - + 2,560 + SK-MEL-177 SK-MEL-177* 1 + 256 + 2,560 + SK-MEL-189 + 4096 + 40,000 + MeWo
Astrocytoma SK-MG-1 SK-MG-4 SK-MG-14 U-251-MG SK-MG-3, -11 SK-MG-21, -23 Neuroblastoma
SK-NMC SK-NSH, IMR-32 Leukemia HL-60 K-562 CCRF-HSB-2
_
_
-
+
-
640 + 2,560
_
_
_
+
-
+
_
-
-
+
10,00 -
-
+
32 +
-
-
+
-
40,000
+ 2,560+
_
+
_
+
-
+
+
+
+
-
+
-
-
+
+
+
_-
+
-
+
-_
_
+
_
+
_
_
-
+
-
+
_-
-
+
+
_
+
+
DSM1 32-27M T A T A
32 + 2,560 + 128 + 40,000 + 4096 + 40,000 + + 10,000 + +
_
-
+ 10,000 + 40,000 +
-
+
_
40,000 +
-
_
+
_-
8+ 8
-
_ _
-
+
_
-
+
_-
-
+
_
+
___-
_
+
_
+
or +
4+
-
_
_
-
_
+
_
_
-
CCRF-CEM, T-45 NALL-1 NALM-1 ARA-10 BALL-1
8+
640 6,400
-
_
FCM1 T A
(1987)
Fibroblasts WI-38 Adult skin Melanocytes (n = 3)
-
-
-
-
-
+
64 + 10,000 + 256 + 2,560 + 4096 or or 2,560 + - or 128
- Normal kidney +- DAUDI, SK-LY-16 epithelium (n = 3) 40,000 EBV-transformed 13 Erythrocytes + + cell A, B, AB, 0 - - AH, DS, HJ Xenogenic cells + + FC - + JB-RH +- 4+ - + - GX 512 + Sheep erythrocytes 2.39M was tested by the same cell panel used here, and it only reacted with sheep erythrocytes (titer, 1:128). T, titer: dilution (reciprocal) showing 50% erythrocyte-rosetted target cells. -, No reaction in direct tests at a dilution of 1:2; ±,
