Znt. J. Cancer: 51,225-231 (1992) 0 1992 Wiley-Liss, Inc.
Publication of the International Union Against Cancer Publicationde I'Union Internationale Contre le Cancer
GLYCOLIPIDS CARRYING Ley ARE PREFERENTIALLY EXPRESSED ON SMALL-CELL LUNG CANCER CELLS AS DETECTED BY THE MONOCLONAL ANTIBODY MLuCl Flavio LE0NI's3, Maria I. COLNAGH11,4,Silvana CANEVARI', Sylvie MENARDI,Elsa COLZANrl, Patrizia FACHERIS', Mariangela FIGINI~, Silvia MIOTTI~ and John L. MAGNANI~ 'Division of Experimental Oncology E, Istituto Nazionale per lo Studio e la Cura dei Tumori, f i a G. Venetian 1, 20133 Milan, Italy; 2BioCarbInc., 300 Professional Drive, Gaithersburg, MD 20879, USA. The monoclonal antibody MLuC I, which reacts strongly with a high percentage of small-cell lung cancers (SCLC), as well as with various human carcinomas, has been used to immunochemically characterize the recognized epitope (CaMLuC I). To this aim 3 different approaches were adopted: (I) immunoblotting/ immunostainingof extracts from various tumor-cell lines; (2) inhibition of binding by purified oligosaccharides; (3) direct binding to oligosaccharide-protein conjugates. All of these experiments indicate that CaMLuC I is present on the Ley blood-group structure heterogeneously expressed on various glycoproteins and glycolipids. The expression of the glycoconjugates carrying Ley was then analyzed on breast and lung cancers and on their normal counterparts. Our overall results suggest that SCLC produce Ley-active glycolipids in higher amounts compared to other tumors of the same or of a different oncotype, as well as normal lung cells, thus indicating an SCLC-specificmodification of the glycosylation pathways.
o 1992 Wiley-Liss,Inc.
Many MAbs developed against human carcinomas recognize oligosaccharide epitopes carried on different glycoconjugates (Magnani, 1987; Hakomori, 1990). The defined epitopes are often correlated to BGA (Feizi, 1985; Clausen and Hakomori, 1989). By exploiting these MAbs, it has been possible to observe BGA in their mature or precursor forms in a number of human tissues, as well as the marked change in their expression which occurs during development, differentiation and neoplastic transformation (Feizi and Childs, 1987; Clausen and Hakomori, 1989; Hakomori, 1989). However, although a large number of reports regarding BGA expression on carcinomas of different histotypes has been accumulated (Blaszczyk et al., 1984; Sakamoto et al., 1986; Clausen and Hakomori, 1989; Feickert et al., 1990), only limited information is available concerning their expression or biosynthesis in SCLC (Rosen et al., 1984; Hirohashi et al., 1988; Zenita et al., 1988). We have previously reported the histopathological characterization (Agresti et al., 1988) of the MAb MLuCl which was found to be helpful in detecting SCLC in bonemarrow samples (MCnard et al., 1988). Analysis of the MLuCl reactivity on cell lines and the tissue distribution of the recognized epitope suggested a possible identity with BGA. In order to verify this hypothesis, immunochemical studies were performed. In addition, the biochemical reactivity of MLuCl with breast and lung tumors and their normal corresponding tissues was analyzed. MATERIAL AND METHODS
Material The murine MAb MLuCl is an IgG2, whose production and characterization have been described (Agresti et al., 1988). Briefly, the MAb reacted with normal and tumor epithelia of the respiratory and digestive organs and with carcinomas of the urogenital system. MLuCl was purified from the ascitic fluid of nu/nu mice (CD1 or BALB/c) by protein A-chromatography (Ey et al., 1978). Radiolabelling was performed by 1251-lactoperoxidase-catalyzed iodination. The human cell lines were all obtained from the ATCC (Rockville, MD), except for N592, SW626 and POVD, which
were kindly provided respectively by Dr. Minna (Naval Institute, Rockville, MD), the Memorial Sloan-Kettering Cancer Center (New York) and Dr. Pratesi (Istituto Nazionale Tumori, Milan). The cell lines were maintained in RPMI 1640 containing 10% FCS. Pure Le-active ceramides were kindly supplied by Dr. V. Ginsburg (NIH, Bethesda, MD) or were prepared as described (Magnani et al., 1987). Purified oligosaccharides and purified standard gangliosides were obtained from BioCarb (Lund, Sweden). The neoglycoproteins (BioCarb, Gaithersburg, MD) shown in Table I, were synthetic glycoproteins which contained 10-20 moles of purified oligosaccharide covalently coupled per mole of BSA. Three different chemical spacer arms were used to couple the oligosaccharides to proteins: p-aminophenyl (PAP), aminophenylethyl (APE) and acetylphenylene (APD). Poly (isobutyl methacrylate) beads were obtained from Polysciences (Warrington, PA). All the other reagents were of analytical grade.
Immunoblotting of extracts separated by SDS-PAGE Solubilized extracts of tumor-cell lines were obtained as described (Leoni et al., 1986). Briefly, cells were washed extensively with RPMI1640 without FCS and then incubated in 50 mM Tris-HC1 buffer, pH 7.4, containing 1% Nonidet P40 and 1 mM PMSF for 20 min at 4°C. The samples were centrifuged at 10,000 g for 15 min and the pellets were discharged. The protein content of the supernatant was determined according to Lowry et al. (1951). Soluble extracts from normal and tumor specimens were obtained as described by Miotti et al. (1989). Briefly, 6 frozen tissue slices of 20 pm (2-3 mg of tissues) were obtained by microtome and treated with 100 p1 of Laemmli (1970) final sample buffer for 60 min at room temperature with continuous stirring, then boiled for 10 min at 100°C. These samples were then centrifuged as above and the pellets were discharged. Equal amounts of extracts from cell lines or tissues (about 300 pg of proteins) were separated by SDS-PAGE and then immunoblotted, as described by Leoni et al. (1986). 3Present address: Laboratorio di Immunologia Cellulare, Centro Ricerche Italfarmaco, Via dei Lavoratori 54,20092 Cinisello Balsamo, Milan, Italy.
4Towhom correspondence and reprint requests should be sent. Abbreviations; BGA, blood-group antigens; ELISA, enzyme-linked immunosorbent assay; FCS, fetal calf serum; HPTLC, high-performance thin-layer chromatography; ICso, 50% inhibitory concentration; IF, immunofluorescence; IPX, immunoperoxidase; MAb, monoclonal antibody; M,, molecular weight; PMSF, phenylmethylsulfonilfluoride; SCLC, small-cell lung cancer; SDS-PAGE, sodium dodecyl sulfatepolyacrylamide gel electrophoresis; THF, tetrahydrofurane; BSA, bovine serum albumin.
Received: December 23,1991.
226
LEON1 E7'AL. TABLE I
C ARROHYDRATE STRUCTURE OF r H E NEOGLYCOPKOTEINS USED FOR ISPITOPE STRUCTURE
DEFINITION
Group I:
Lacto-N-fucopentaose I (LNF I) Fuc(al-2)Gal( p1-3)NAc( p1-3)Gal( pl-4)Glc-APD Lacto-N-fucopentaose I1 LNF 11) Gal( p1-3)GlcNAc( pl-3\Gal( 61-4)Glc-APD 4
I
Fucoll Lacto-N-fucopentaose 111 (LNF 111) Gal( pl-4)GlcNAc(p 1 -3)Gal( p1-4)Glc-APD 3
I
Fucal Lacto-N-difucohexaose 1 (LND I)
Fuc(oll-2)Gal(pl-3)GIcNAc(pl-3)Gal( p1-4)Glc-APD 4 I
Fucal Lacto-N-difuconeohexaose (LNnD I) Fuc(al-2)Gal( pl-4)GlcNAcpl-0-APE 3
Group 11:
Fucdil Maltose GIC(CY~-~)GICCX~-O-PAP Lactose Gal( p1-4)Glcp 1-0-PAP Lacto-N-tetraose (LNT) Gal( p l-3)GlcNAc( p 1-3)Gal(pl-4)Glc-APD Lacto-N-neotetraose (LNnT) Gal( pl-4)GlcNAc( pl-3)Gal(p1-4)Glc-APD Lacto-N-hexaose (LND)
Gal(~1-4)GlcNAc(pl-6)[Gal(pl-4)GlcNAc(pl-3)]Gal(~1-4)Glc-APD
Group 111:
Lacto-N-neohexaose (LNnD) Gal( p 1-4)GlcNAc(p 1-6)[GaI(pl-3)GlcNAc(pl-3)]Gal( p1-4)Glc-APD Melibiose Gal(a1-6)Glcpl-0-PAP Cellobiose Glc(P 1-4)Glcpl-O-PAP A-trisaccharide
GalNAc(al-3)[Fuc(al-2)]Galpl-O-APE B-trisaccharide Gal(a 1-3)[ Fuc(al-2)]Galpl-O-APE A-tetrasaccharide GalNAc(al-3)[Fuc(al-2)jGal(p1-4)Glc-APD A-heptasaccharide
GalNAc(al-3)[Fuc(a1-2)]Gal( pl-3)[Fuc(a1-4)]GlcNAc(pl-3)Gal(p1-4)GlcAPD H type 2
Fuc(al-2)Gal(pl-4)GlcNAcpl-O-APE Gangliote traose
Gal(pl-3)GalNAc(pl-4)Gal(pl-4)Glc-APD Group IV:
T-antigen Gal@ 1-3)GalNAcal-0-APE 3' Sialyllactose NeuSAc(a2-3)Gal( p1-4)Glc-APD 6' Sialyllactose
NeuSAc(a2-6)Gal(pl-4)Glc-APD LS-tetrasaccharide a NeuSAc(a2-3)Gal( p l-3)GlcNAc( p 1-3)Gal(pl-4)Glc-APD LS-tetrasaccharide b Gal( pl-3)[ (Neu5Ac(a2-6)]GlcNAc(pl-3)Gal(pl-4)Glc-API~ LS-tetrasacharide c NeuSAc(a2-6)Gal( p1-4)GlcNAc( p 1-3)Gal( p1-4)Glc-APD Disialylated lacto-N-tetraose Neu5Ac(a2-3)Gal(pl-3)[NeuSAc(ol2-6)]GlcNAc(~l-3)Gal( p1-4)Glc-APD Sialylated lacto-N-fucopentaose I1 (SLNF 11) NeuSAc(ol2-3)Gal(p 1-3)[Fuc(al-4)]GlcNAc(p1-3)Gal( pl-4)Glc-APD Sialylated lacto-N-fucopentaose I11 (SLNF 111) NeuSAc(a2-3)Gal( pl-4)[Fuc(al-3)]GlcNAc(p1-3)Gal( pl-4)Glc-APD
REACTION OF MLuCl ANTIBODY WITH THE LEY DETERMINANT
227
TABLE I (continued)
Group V: Biantennary octasaccharide Gal( p 1-4)GlcNAc(pl-2)Man(a1-6)[Gal( p 1-4)GlcNAc(p1-2)Man(a1-3)] Man@ 1-4)GlcNAc-APD Globotriose Gal(a1-4)Gal( p1-4)Glc-AF'D Globotetraose GalNAc( p 1-3)Gal(al-4)Gal( p1-4)Glc-APD Man2GlcNAc Man(al-3)Man(PI-4)GlcNAc-APD Man3GLcNAc Man(a1-2)Man(a1-3)Man( p1-4)GlcNAc-APD Chitotriose GIcNAc(~~-~)GIcNAc( pl-4)GkNAc-APD G r o w VI: Bovine serum albumin (background control)
Abbreviations:Fuc, fucose; Gal, galactose; Glc NAc, N acetylglucosamine; Glc, glucose; Gal NAc, N acetyl galactosamine; Neu 5Ac, neuraminic acid; Man, mannose. Immunostaining of high-performance thin-layer chromatograms of glycolipids Glycolipid extracts from cell lines and from normal and tumor specimens were obtained by THF extraction (Leoni et al., 1986; Magnani et al., 1987). The total glycolipid extracts and the purified Le-active ceramides (2 pg) were dried by evaporation under nitrogen stream, then dissolved in methanol and applied to silica HPTLC plates. After chromatography in chloroform/methanol/0.25% KCI (5:4:1) the plates were soaked for 1 min in 0.1% of poly(isobutylmethacry1ate) beads dissolved in hexane and immunostained by '2SI-labelledMLuCl (10' cpm/ml) for 60 min at room temperature. After extensive washing with phosphate buffer, the plates were submitted to autoradiography (Leoni et al., 1986). Inhibition of MLuCl-binding by purified oligosaccharides The effect of purified oligosaccharides on the binding of MLuCl to a total glycolipid extract of A431 cells was tested by solid-phase radioimmunoassay carried out on polyvinylchloride microtiter plates coated with the extract (Leoni et al., 1988). The amount of A431 extract was determined by a previous titration with '251-labelled MLuC1. Briefly, serial dilutions of purified oligosaccharides were incubated with 1251-labelledMLuCl (5 x lo5 cpm/well) for 1 hr at room temperature. The mixture was then transferred to glycolipidcoated plates and incubated for 2 hr. After washing with phosphate buffer the wells were cut from the plates and the bound radioactivity was determined with a gamma counter. The percentage of MLuCl binding-inhibition exerted by each oligosaccharide dilution was calculated. Direct binding of MLuCl to neoglycoproteins The binding of MLuCl to 34 different neoglycoproteins was tested by a conventional ELISA. The carbohydrate structures on the neoglycoproteins are shown in Table I. To simplify the screening procedure, the 34 different neoglycoproteins were divided into 5 groups. Wells coated with a balanced mixture of glycoproteins from each group (total amount 100 pgiwell) were used for initial screening. Each member of any positive group was then individually assayed at a concentration of 20 pg/well. Immunohistochemistly Surgical specimens were obtained from cancer patients. Pathological or normal adjacent tissues were freed of necrotic areas as far as possible and frozen in nitrogen. MLuCl reactivity on tumor-cell lines and on tissues sections was evaluated by IF and/or IPX (Tagliabue et al., 1985; Agresti et al., 1988).
FIGURE1 - Binding reactivity of '2sI-labelled MLuCl on: (a) materials separated by SDS-PAGE and blotted on nitrocellulose after solubilization from MCF7 (l), A431 (2) and SW626 (3) tumor-cell lines; (b) HPTLC-separated total glycolipid extracts from N592 (4), MCF7 ( 5 ) and A431 (6) tumor cell lines. Nitrocellulose paper and HPTLC plates were autoradiographed for 20 hr at -70°C in the presence of an intensifying screen. Molecular weight markers for SDS-PAGE are indicated on the right of panel (a)' RESULTS
Characterization of MLuCl antigenic molecules Single-cell suspensions from various human tumor-cell lines were evaluated by I F to determine the best antigenic source for the biochemical characterization of the antigen. MLuCl reacted with 11 out of 15 cell lines of epithelial origin, whereas none of the 4 cell lines of non-epithelial origin (melanoma, sarcoma and lymphoma) were reactive (data not shown). Three epithelial cell lines, i.e. MCF7 (breast carcinoma), A431 (vulvar carcinoma) and SW626 (ovarian carcinoma) were used in the following biochemical analysis. Solubilized extracts were analyzed by immunoblotting with MLuCl and the relevant results are reported in Figure la. MLuCl strongly labelled materials of heterogeneous M, (ranging approximately from less than 14,000 to over 200,000), with a unique staining pattern for each sample. On the basis of the apparent M, and broad migration, the MLuC1-positive molecules, which excccded 14,000, were considered to be glycoproteins. The
228
LEON1 b7 AL.
samples from A431 and SW626 showed a well-defined reactive band which migrates with the ion boundary. According to our previous data (Leoni et al., 1986), this band strongly suggests the presence of antigenic glycolipids. Total glycolipid extracts, obtained from 2 of the previous lines and from the N592 cell line (SCLC), were separated by HPTLC and immunostaincd by MLuC1. As depicted in Figure lb, several immunoreactive bands were detected on all 3 extracts, including that from the MCF-7 linc (lane 5), thus suggesting that the expression of the MLuCl -positive glycolipids was below the threshold value for detection by immunoblotting.
Characterization of the carbohydrate epitope recognized by MLuC1 The epitope recognized by MLuCl was characterized by 3 different independent experimental approaches. In the first, titratcd doses of purified oligosaccharides from human milk were used to inhibit thc binding of '2'I-labelled MLuCl to glycolipids extracted from the A431 tumor cell linc. As shown in Figure 2, only lacto-difucotetraose (LDFT, Ley-like hapten) was able to totally inhibit the binding and gave an IC,, of about 800 pM. LNF I1 (Lea-hapten) exerted only a slight inhibition (around 25%) at the maximum concentration tested, whereas LNF 111 (Le"-hapten), LND I (Le"-hapten) and lactose were completely ineffective. For the second experimental approach, direct binding of "'I-labelled MLuCl to purified Le-active glycolipids was carried out by immunostaining HPTL chromatograms. As shown in Figure 2 (insert), the Ley-ceramide was strongly stained, whereas the Le", Le" and Le"-ccramides wcrc completely negative. For the final approach, MLuCl was assayed by ELISA for direct binding to a panel of 34 different neoglycoproteins
FIGURE 2 - Inhibition of the binding of i2sI-labelledMLuCl by oligosaccharides. Total glycolipid extract from the A431 cell line was used as antigenic source. MLuCl binding was evaluated in the presence of the same concentrations of LDFT (y-like), LNFIT (a), LNFIII (x), LNDI (b) and lactose. For clarity the experimental points are reported only for LDFT. For the structure of the oligosaccharides see Table I:
LDFT = Fuc(al-2)Gal(B i -4)Glc 3
i
Fuca 1 Insert: binding of 12iI-labelled MLuCl on HPTLC-separated purified Ley, LeX, Lea and Leh ceramides. The more slowly migrating band present in the Ley-ceramide preparation, faintly stained by MLuC1, could be attributed to the presence of glycolipids with extended saccharide moieties.
divided into 5 groups, as shown in Table 1. MLuCl only bound to group I (Fig. 3a) and when each neoglycoprotein in group I was assayed individually, only the Ley-active neoglycoprotein (LNnD I) was reactive (Fig. 3b).
Expression of the CaMLuCl -canying glycolipids on breast tissues Since both normal and cancerous breast epithelium were positive with MLuCl (Agresti et al., 1988) the expression of CaMLuCl was further analyzed on surgical specimens by IPX and by immunoblotting, and the relevant results are reported in Table 11. By IPX normal ductal epithelial cells from mammary glands (either resting or lactating) were diffusely positive at both cytoplasmic and membrane levels in more than 60% of the tested cases. With rcgard to malignant breast tumors, about 75% of the tested cases showed either focal membrane or diffused membrane and cytoplasmic staining. By immunoblotting, a limited number of samples from rmting or hormone-stimulated normal tissues were analyzed. Out of the 4 normal and the 5 benign samples tested, 3 normal and 2 benign samples showed immunostained glycoproteins of different M, but in no case were glycolipids identified. As regards the 30 samples from carcinomas, 11 MLuCl-positive cases were found and the majority of the reactive molecules were glycoproteins. However, in 2 cases a glycolipid band was also clearly detectable. Expression of the CaMLuCI-canying glycolipids on lung tissues Since the antibody MLuCl appears to be of clinical interest in the case of SCLC (MCnard et al., 1988),we further evaluated its reactivity on normal and tumorous lung tissues. As reported in Table 111, a strong expression of CaMLuCl was observed on all the cell lines derived from SCLC, whereas the 3 lung cancer lines of different histotypes were found to be negative. At variance with the cell linc data, the majority of the specimens taken from the lung tumors were positive regardless of their histotypes. The staining pattern was, however, different, being homogeneous on SCLC and adenocarcinomas and hctcrogeneous on squamous carcinomas. Moreover, as previously reported (Agresti et al., 1988), a focal reactivity was observed on normal bronco-alveolar glands of about half of the normal lung specimens tested. Analysis of the glycoconjugates responsible for the MLuCl reactivity was performed by 2 different approaches. Solubilized extracts from 3 SCLC cell lines and a few available specimens of normal and tumorous lung tissues were analyzed by immunoblotting with MLuC1. As shown in Figure 4, on all but one (lane c ) of the SCLC samples, either from cell lines or from surgical specimens, MLuCl seemed to preferentially or exclusively identify glycolipid molecule(s), whereas on one squamous (lane g) and one adeno (lane h) carcinoma sample a faint staining was observed only on glycoproteins of different M,. According to the focal reactivity on normal tissues and to the low sensitivity of the method, no staining was detected on the 2 normal lung specimens tested (data not shown). According to cell-line analysis (see Fig. 1), the immunostaining of HPTLC-separated glycolipid extracts appeared to be more sensitive than immunoblotting in detecting the presence of MLuCl -reactive glycolipids. Therefore, in a selected series of individual patients, the expression of CaMLuCl was evaluated on the same surgical specimens by I F on frozen sections and by immunostaining on the total glycolipid extracts. As reported in Table IV, MLuCl weakly stained by IF about 10% of cells in 2 out of 4 normal lung specimens, whereas no reactivity was seen with the 4 glycolipid extracts. Similarly, all the samples except one from the 7 adeno and squamous carcinomas tested were strongly positive by IF, whereas in only 1 out of the 7 samples were MLuCl-reactive glycolipids detected.
229
REACTION OF MLuCl ANTIBODY WITH THE LEY DETERMINANT o.8 Absorbance at 650 nm -
~
_
_
_
~
~
7 _i _
1
Absorbance at 650 nm
E 0.6
0.6
d.
0.4
0.2
0 I
0.4
0.2
V
IV
111
II
0.6
BSA
0 LNnDl
Neoglycoprotein group
LNFl
LNFll
LNFlll
LNDl
Maltose
BSA
FIGURE 3 - MLuCl binding by ELISA to: (a) neoglycoprotein groups; (b) each neoglycoprotein of group I. For the structure of the oligosaccharides see Table I. TABLE I1 - MLuCl BINDING REACTIVITY ON NORMAL AND TUMOR BREAST TISSUES AS DETERMINED BY IMMUNOHISTOCHEMISTRY AND IMMUNOBLOlTING
Breast tissues
Normal Benigntumor Carcboma3
Number of surgical specimens
11
5 30
Number positiveinumber tested Immunohistochemistm' F
D
Oil1 7/11 ND5 ND 11/29 11/29
lmmunoblotting* Glycoproteins Glycolipids
314 215 11/30
014
015 24/30
'On frozen section by IPX. Intensity and percentage of positivity expressed as: F = focally positive on less than 30% of cells; D = diffusely positive on more than 70% of cells.-zOn tissue extracts. Immunostained molecules were operationally defined as: glycoproteins when separated in the separation el and glycolipids when migrated with the boundary of ions.%xAtding primary and meta~tasis.-~Thepositive cases also expressed glycoproteic molec~les.-~ND, not determined.
In contrast, in the case of SCLC specimens, good agreement could be observed between the results of the 2 tests (in 3 out of the positive cases by IF the glycolipid antigen was identified). In all the positive cases the HPTLC staining pattern was very similar to that shown in Figure l b for the cell lines. DISCUSSION
The epitope recognized by the MLuCl MAb was analyzed by 3 different independent immunochemical methods. All of the results demonstrated that MLuCl binds specifically only to the Ley-active tetrasaccharide structure. The MAb therefore seems to differ from some of the already described anti-Ley antibodies which showed a cross-reactivity with other similar structures (Fukushi et al., 1984; Gooi et al., 1985). The wide binding analysis performed on different oligosaccharides enables us to define the saccharide sequence recognized by the MAb. The results indicate that the N-acetyl group of N-acetylglucosamine is not required for MAb binding; in fact, LDFT (see legend of Fig. 3 for the structure), which contains glucose as the reducing sugar, is also able to inhibit MLuCl binding. Therefore, the proposed epitope is a type-2 chain core with 2 fucosyl substitutions. The pattern of reactivity of MLuCl with various human carcinomas and particularly with SCLC (Agresti et al., 1988) is in agreement with some (Sakamoto et al., 1986; Hirohashi et al., 1988) but not all (Zenita et al., 1988) of the reports concerning anti-Ley MAbs. This discrepancy may be due to slight variations in the epitope or in the binding affinities of the various antibodies.
TABLE 111 MLuCl IMMUNOHISTOCHEMICAL REACTIVITY ON CELL LINES FROM LUNG TUMORS AND ON NORMAL AND TUMOROUS LUNG TISSUES ~
Material Cell lines
SCLC NCI-H69 NCI-H128 N592 POVD Adenocarcinoma CaLu3 AS49 Squamous carcinoma SKMESl
MLuCl reactivitv Intensitv'
++ ++ ++ ++
% of positive cells
95 90
90 95 -
-
-
Surgical specimens
Normal SCLC Adenocarcinoma Sauamous carcinoma
316 45/52 17/21 18124
'On live cells by IF + + = homogeneous staining.-20n frozen sections by IF. In brackets, the intensity and % of positivity expressed as: F = focally positive on less than 30% of cells; H = heterogeneously positive (from 30 to 70% of cells); D = diffusely positive on 70-100% of cells. The difference in the fine specificity of the antibodies could probably also account for the different pattern of reactivity exhibited by MLuCl on A431 compared to the previously reported data (Basu et al., 1987). Analysis of the glycoconjugates which carry the Ley-related epitope recognized by MLuCl was mainly performed by immunoblotting since this procedure was previously found to be suitable for simultaneous identification of the presence of both glycoproteins and glycolipids (Leoni et al., 1986). However, its exploitation was limited by the level of sensitivity which, if compared to immunohistochemistry, was quite low. Therefore, in the case of SCLC, immunostaining of HPTLCseparated glycolipids was also applied. The analysis, although performed in a restricted number of cases and within the limits of the applied methodologies, suggested that SCLC, as compared to other lung and breast tumors and their normal corresponding tissues, preferentially exhibited glycolipid structures. This observed overexpression of Ley glycolipids on SCLC could account for the apparent discrepancy between the high frequency of increased Ley level seen in the serum of patients
230
LEON1 E T A L . TABLE IV
MLuCl REAC IIVITY ON SELECTED CASES OF NORMAL AND TUMOROUS LUNG TISSUES AS DETERMINED BY IMMUNOFIISTOCFIEMISTRY AND IMMUNOSTAINING ~
Number posilivelnumber tested
Lung tissues
lmmunohistochemistry
HPTLC immunoreaction*
Normal Tumor adenocarcinoma squamous carcinoma SCLC
214 (F)
014
:;: {El
113 014 314
414 (D)
'On frozen sections by IF. In brackets, the intensity and % of positivity expressed as: F = focally positive on less than 30% of cells; H = heterogeneously positive (from 30 to 70% of cells); D = diffusely positive on 70-100% of cells.-20n total glycolipid extract.
FIGURE 4 - Binding reactivity of 'Z'I-labelled MLuCl on material separated by SDS-PAGE and blotted on nitrocellulose after solubilization from: H69 (a), N592 (b), POVD (c) SCLC cell lines and from surgical specimens of 3 SCLC (d, e and f), 1 squamous (8) and 1 adeno (h) carcinoma of the lung. The paper was autoradiographed for 168 hr at -70°C in thc presence of an intensifying screen. Molecular weight markers are indicated on the right. with different carcinomas (Kannagi et al., 1986), and the low frequency observed in lung-cancer patients. Since B G A a r e mainly found in the serum on large mucin-like molecules (Magnani et al., 1983), it is possible that the preferential expression on glycolipids in SCLC prevents their shedding. In the case of SCLC the overcxprcssion of other glycolipids such as the hydroxy fucosyl-GM1 (Nilsson et al., 1986), GD2 (Cheresh et al., 1986) and the glycolipid recognized by the antibody SMI (Bernal et al., 1988) has already been described. T h e genetic alterations which lead to transformation are gencrally accompanied by modulation of the biosynthesis of
cell-surfacc glycoconjugates. As a consequence, the aberrant expression of glycolipid structures on tumor cells has been reported for several cancers (Hakomori, 1985). T h e glycolipids which are neosynthetized or overexpresscd by tumor cells may play a role in several processes such as cell attachment t o the matrix (Hakomori, 1981), growth regulation and response t o growth factors (Bremer et al., 1984). Thcrefore, the MLuCl M A b may b e particularly useful in the case of SCLC, where glycolipids carrying Ley are extremely abundant, to examine the mechanism of aberrant glycosylation in tumor cells and its effect o n the functions of cell-surface glycoconjugates. ACKNOWLEDGEMENTS
This work was partially supported by a grant from the Associazione per la Ricerca sul Cancro. W e thank Ms. G. Ferrami, Ms. C. Andrews and Ms. Seiler for excellent technical assistance, Ms. L. Mameli and Ms. M. Hatton for manuscript prcparation and Mr. M. Azzini for the photographic reproductions.
REFERENCES
Cell-surface antigens of human lung tumors detected by mouse AGRESTI,R., ALZANI,R., ANDREOLA, S., BEDINI,V., G I A N ~S., , MENARD, S., RILKE,F. and COLNAGHI, M.I., Histopathological charac- monoclonal antibodies: definition of blood-group and non-blood-groupterization of a novel monoclonal antibody, MLuCl, reacting with lung related antigenic systems. Int. J. Cancer, 46,1007-1013 (1990). carcinomas. Tumor;, 74,401-410 (1988). FEIZI,T., Demonstration by monoclonal antibodies that carbohydrate BASU,A,, MUKTHY, U., RODECK,U., HERLYN,M., MATTES, L. and structures of glycoproteins and glycolipids are onco-developmental DAS, M., Presence of tumor-associated antigens in epidermal growth antigens. Nature (Lond.), 314,53-57 (1985). factor receptors from different human carcinomas. Cancer Res., 47, FEIZI,T. and CHILDS,R.A., Carbohydrates as antigenic determinants 2531-2536 (1987). of glycoproteins. Biochem. J., 245,l-I1 (1987). BERNAL,S.D., CUALING, H.M., ELIAS,A D . , SCHWAKWNG, G. and FUKUSHI, Y . ,HAKOMOKI, S., NUDELMAN, E. and COCHRAN, N., Novel MCCLUER,R.M., Classification of small cell carcinoma antigens by fucolipids accumulating in human adenocarcinoma. 11. Selective isolatheir association with subcellular components. Lung Cancer, 4,99-102 tion of hybridoma antibodies that differentially recognize mono-, di-. (1988). and trifucosylated type 2 chain. J. bid. Chem., 259,4681-4685 (1984). BLASZCZYK, M., HANSSON, G.C., KARLSSON, K.A., LARSON, G., STROM- GOOI,H.C., PICAKD, J.K., H O U N S ~ LE.F., L , GREGOKIOU, M., R ~ E s , BERG, N., THURIN, J., MEENHAKD, H., STEPLEWSKI, Z. and KOPKOWSKI, A.R. and FEIZI,T., Monoclonal antibody (EGFIG49) reactive with the H., Lewis blood group antigen defined by monoclonal anti-colon epidermal growth factor receptor of A431 cells recognizes the blood carcinoma antibodies. Arch. hiochem. Riophyr., 233,161-168 (1984). group Aleh and Ale' structures. Mol. Immunol., 22,689-693 (1985). BREMER,E.G., HAKOMORI, S.I., BOWEN-POPE, D.F., RAINES,E. and HAKOMORI, S., Glycolipids in cellular interaction, differentiation, and Ross, R., Ganglioside-mediated modulation of cell growth, growth factor binding, and receptor phosphorylation. J. biol. Chem., 259, oncogenesis. Ann. Rev.Biochem., 50,733-764 (1 981). HAKOMORI, S., Aberrant glycosylation in cancer cell membranes as 6818-6825 (1984). on glycolipids: overview and perspectives. Cancer Res., 45, CHEKESH.D.A., R O S E N B ~ KJ., G ,MUJOO,K., HIRSCHOWI-IL., L. and focused REWELD,R.A., Biosynthesis and expression of the disialoganglioside 2405-2414 (1985). S., Aberrant glycosylation in tumors and tumor-associated GD2, a relevant target antigen on small cell lung carcinoma for HAKOMORI, carbohydrate antigens.Advanc. CuncerRes., 52,257-331 (1989). monoclonal antibody-mediated cytolysis. Cancer Rex, 46, 5112-5118 (1986). HAKOMORI, S., Biochemical basis of tumor-associated carbohydrate CLAUSEN, H. and HAKOMORI, S., ABH and related histo-blood-group antigens. Immunol. Allergy Clin. N. Amer., 4,781-802 (1990). antigens: immunochemical differences in carrier isotypes and their HIKOHASHI, S., KUNII,T., HIZANO,T. and SHIMOSAI'O, Y., 145-kDa distribution. Vox Sang., 56, 1-20 (1989). cells membrane antigen on SCLC as a common target for several EY.P.L., PROWSE,S.L. and JENKIN, C.R., Isolation of pure IgG,, IgG, monoclonal antibodies raised against SCLCs. Lung Cancer, 4,103-104 and IgGZ, immunoglobulins from mouse serum using prutein (1988). A-Sepharose. Inimunochemisty, 15.429-436 (1978). KANNAGI, R., FLJKUSHI. Y., TACHIKAWA, T., NODA, A,, SHIN,S., Y., INAMOTO, T., HAKOMORI, S. K., HIKAIW.4, N., FUKUDA, FEICK~RT. H.J., ANGLR,B.R., CORDON-CAKDO. C. and LLOYO,K., SHIGETA,
REACTION OF MLuCl ANTIBODY WITH THE LEY DETERMINANT
and IMURA,H., Quantitative and qualitative characterization of human cancer-associated serum glycoprotein antigens expressing fucosyl or sialyl-fucosyl type 2 chain polylactosamine. Cancer Res., 46, 2619-2626 (1986). LAEMMLI, U.K., Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (Lond.), 227,680-685 (1970). LEONI,F., MAGNANI, J.L., Mion.~,S., CANEVARI, S., PASQUALI, M., S. and COLNAGHI, M.I., The antitumor monoclonal antibody SONNINO, MOv2 recognizes the Lewis A hapten. Hybridoma, 7,129-139 (1988). LEONI,F., MIOITI,S., CANEVARI, S., SONNINO, S., RIPAMONTI, M. and COLNAGHI,M.I., Carbohydrate epitope defined by an antitumor monoclonal antibody detected on glyco roteins and a glycolipid by immunoblotting. Hybridoma, 5,289-296 8986). LOWRY,O.H., ROSEBROUGH, N.J., FARR,A.L. and RANDALL, R.J., Protein measurement with the Fohn phenol reagent. J. bid. Chem., 193,265-275 (1951). MAGNANI, J.L., Mouse and rat monoclonal antibodies directed against carbohydrates. Meth. EnzymoL, 138,484491 (1987). MAGNANI, J.L., SPITALNIK, S.L. and GINSBURG, V., Antibodies against cell surface carbohydrates: determination of antigen structure. Meth. Enzymol., 138,195-207 (1987). MAGNANI, J.L., STEPLEWSKI, Z., KOPROWSKI, H. and GINSBURG, V., Identification of the gastrointestinal and pancreatic cancer-associated antigen detected by monoclonal antibody 19-9 in the sera of patients as a much. Cancer Res., 43,5489-5492 (1983). M~NARD S.,, PORRO,G., GIAN‘L, S., BEDINI, A,, BIDOLI,P., SANTORO, A., BREGNI,M., GIANNI,A.M., PiLoi-rI, S., RILE, F., SORESI,S., FASOLATO, S. and COLNAGHI, M.I., Detection of SCLC metastatic cells
23 1
in bone marrow by means of the monoclonal antibody MLuCl. Lung Cancer, 4,73-75 (1988). MIOTTI,S., LEONI,F., CANEVARI, S., SONNINO, S. and COLNAGHI, M.I., Immunoblotting detection of carbohydrate epitopes in glycolipids and glycoproteins of tumoral origin. In: H.F. Oettgen (ed.), Gangliosides and cancer, pp. 168-176, VCH Publ., Weinheim (Germany) (1989). NILSSON, O., BREZICKA, F.T., HOLMGREN, J., SORENSON, S., SVENNERF. and LINDHOLM, L., Detection of a ganglioside HOLM, L., YNGVASON, antigen associated with small cell lung carcinomas using monoclonal antibodies directed against fucosyl-GM1. Cancer Rex, 46, 1403-1407 (1986). J.L., CUTTITTA,F., FEDORKO, J., CARNEY, ROSEN,S.T., MULSHINE, D.N., GAZDAR, A.F. and MINNA,J.D., Analysis of human small cell lung cancer differentiation antigens using a panel of monoclonal antibodies. Cancer Res., 44,2052-2061 (1984). J., FURUKAWA, K., CORDON-CARDO, C., YIN, B.W.T., SAKAMOTO, RETTIG,W.J., OETTGEN, H.F., OLD,L.J. and LLOYD,K.O., Expression of Lewis”, Lewisb, X and Y blood group antigens in human colonic tumors and normal tissue and in human-derived cell lines. CancerRes., 46,1553-1561 (1986). E., MENARD,S., DELLATORRE,G., BARRANTI, P., TAGLIABUE, MARIANI-COSTANTINI, R., PORRO,G. and COLNAGHI, M.I., Generation of monoclonal antibodies reacting with human epithelial ovarian cancer. Cancer Res., 45,379-385 (1985). A., SHIGETA, K., HIGUCHI, K., ZENITA,K., KIRIHATA, Y., KITAHARA, HIRASHIMA, K.;MURAKI,MIYAKE, M., TAKEDA, T. and KA”AGI, R., Fucosylated type-2 polylactosamine antigens in human lung cancer. Int. J. Cancer, 41,344-349 (1988).
Publication of the International Union Against Cancer Publicationde I'Union Internationale Contre le Cancer
GLYCOLIPIDS CARRYING Ley ARE PREFERENTIALLY EXPRESSED ON SMALL-CELL LUNG CANCER CELLS AS DETECTED BY THE MONOCLONAL ANTIBODY MLuCl Flavio LE0NI's3, Maria I. COLNAGH11,4,Silvana CANEVARI', Sylvie MENARDI,Elsa COLZANrl, Patrizia FACHERIS', Mariangela FIGINI~, Silvia MIOTTI~ and John L. MAGNANI~ 'Division of Experimental Oncology E, Istituto Nazionale per lo Studio e la Cura dei Tumori, f i a G. Venetian 1, 20133 Milan, Italy; 2BioCarbInc., 300 Professional Drive, Gaithersburg, MD 20879, USA. The monoclonal antibody MLuC I, which reacts strongly with a high percentage of small-cell lung cancers (SCLC), as well as with various human carcinomas, has been used to immunochemically characterize the recognized epitope (CaMLuC I). To this aim 3 different approaches were adopted: (I) immunoblotting/ immunostainingof extracts from various tumor-cell lines; (2) inhibition of binding by purified oligosaccharides; (3) direct binding to oligosaccharide-protein conjugates. All of these experiments indicate that CaMLuC I is present on the Ley blood-group structure heterogeneously expressed on various glycoproteins and glycolipids. The expression of the glycoconjugates carrying Ley was then analyzed on breast and lung cancers and on their normal counterparts. Our overall results suggest that SCLC produce Ley-active glycolipids in higher amounts compared to other tumors of the same or of a different oncotype, as well as normal lung cells, thus indicating an SCLC-specificmodification of the glycosylation pathways.
o 1992 Wiley-Liss,Inc.
Many MAbs developed against human carcinomas recognize oligosaccharide epitopes carried on different glycoconjugates (Magnani, 1987; Hakomori, 1990). The defined epitopes are often correlated to BGA (Feizi, 1985; Clausen and Hakomori, 1989). By exploiting these MAbs, it has been possible to observe BGA in their mature or precursor forms in a number of human tissues, as well as the marked change in their expression which occurs during development, differentiation and neoplastic transformation (Feizi and Childs, 1987; Clausen and Hakomori, 1989; Hakomori, 1989). However, although a large number of reports regarding BGA expression on carcinomas of different histotypes has been accumulated (Blaszczyk et al., 1984; Sakamoto et al., 1986; Clausen and Hakomori, 1989; Feickert et al., 1990), only limited information is available concerning their expression or biosynthesis in SCLC (Rosen et al., 1984; Hirohashi et al., 1988; Zenita et al., 1988). We have previously reported the histopathological characterization (Agresti et al., 1988) of the MAb MLuCl which was found to be helpful in detecting SCLC in bonemarrow samples (MCnard et al., 1988). Analysis of the MLuCl reactivity on cell lines and the tissue distribution of the recognized epitope suggested a possible identity with BGA. In order to verify this hypothesis, immunochemical studies were performed. In addition, the biochemical reactivity of MLuCl with breast and lung tumors and their normal corresponding tissues was analyzed. MATERIAL AND METHODS
Material The murine MAb MLuCl is an IgG2, whose production and characterization have been described (Agresti et al., 1988). Briefly, the MAb reacted with normal and tumor epithelia of the respiratory and digestive organs and with carcinomas of the urogenital system. MLuCl was purified from the ascitic fluid of nu/nu mice (CD1 or BALB/c) by protein A-chromatography (Ey et al., 1978). Radiolabelling was performed by 1251-lactoperoxidase-catalyzed iodination. The human cell lines were all obtained from the ATCC (Rockville, MD), except for N592, SW626 and POVD, which
were kindly provided respectively by Dr. Minna (Naval Institute, Rockville, MD), the Memorial Sloan-Kettering Cancer Center (New York) and Dr. Pratesi (Istituto Nazionale Tumori, Milan). The cell lines were maintained in RPMI 1640 containing 10% FCS. Pure Le-active ceramides were kindly supplied by Dr. V. Ginsburg (NIH, Bethesda, MD) or were prepared as described (Magnani et al., 1987). Purified oligosaccharides and purified standard gangliosides were obtained from BioCarb (Lund, Sweden). The neoglycoproteins (BioCarb, Gaithersburg, MD) shown in Table I, were synthetic glycoproteins which contained 10-20 moles of purified oligosaccharide covalently coupled per mole of BSA. Three different chemical spacer arms were used to couple the oligosaccharides to proteins: p-aminophenyl (PAP), aminophenylethyl (APE) and acetylphenylene (APD). Poly (isobutyl methacrylate) beads were obtained from Polysciences (Warrington, PA). All the other reagents were of analytical grade.
Immunoblotting of extracts separated by SDS-PAGE Solubilized extracts of tumor-cell lines were obtained as described (Leoni et al., 1986). Briefly, cells were washed extensively with RPMI1640 without FCS and then incubated in 50 mM Tris-HC1 buffer, pH 7.4, containing 1% Nonidet P40 and 1 mM PMSF for 20 min at 4°C. The samples were centrifuged at 10,000 g for 15 min and the pellets were discharged. The protein content of the supernatant was determined according to Lowry et al. (1951). Soluble extracts from normal and tumor specimens were obtained as described by Miotti et al. (1989). Briefly, 6 frozen tissue slices of 20 pm (2-3 mg of tissues) were obtained by microtome and treated with 100 p1 of Laemmli (1970) final sample buffer for 60 min at room temperature with continuous stirring, then boiled for 10 min at 100°C. These samples were then centrifuged as above and the pellets were discharged. Equal amounts of extracts from cell lines or tissues (about 300 pg of proteins) were separated by SDS-PAGE and then immunoblotted, as described by Leoni et al. (1986). 3Present address: Laboratorio di Immunologia Cellulare, Centro Ricerche Italfarmaco, Via dei Lavoratori 54,20092 Cinisello Balsamo, Milan, Italy.
4Towhom correspondence and reprint requests should be sent. Abbreviations; BGA, blood-group antigens; ELISA, enzyme-linked immunosorbent assay; FCS, fetal calf serum; HPTLC, high-performance thin-layer chromatography; ICso, 50% inhibitory concentration; IF, immunofluorescence; IPX, immunoperoxidase; MAb, monoclonal antibody; M,, molecular weight; PMSF, phenylmethylsulfonilfluoride; SCLC, small-cell lung cancer; SDS-PAGE, sodium dodecyl sulfatepolyacrylamide gel electrophoresis; THF, tetrahydrofurane; BSA, bovine serum albumin.
Received: December 23,1991.
226
LEON1 E7'AL. TABLE I
C ARROHYDRATE STRUCTURE OF r H E NEOGLYCOPKOTEINS USED FOR ISPITOPE STRUCTURE
DEFINITION
Group I:
Lacto-N-fucopentaose I (LNF I) Fuc(al-2)Gal( p1-3)NAc( p1-3)Gal( pl-4)Glc-APD Lacto-N-fucopentaose I1 LNF 11) Gal( p1-3)GlcNAc( pl-3\Gal( 61-4)Glc-APD 4
I
Fucoll Lacto-N-fucopentaose 111 (LNF 111) Gal( pl-4)GlcNAc(p 1 -3)Gal( p1-4)Glc-APD 3
I
Fucal Lacto-N-difucohexaose 1 (LND I)
Fuc(oll-2)Gal(pl-3)GIcNAc(pl-3)Gal( p1-4)Glc-APD 4 I
Fucal Lacto-N-difuconeohexaose (LNnD I) Fuc(al-2)Gal( pl-4)GlcNAcpl-0-APE 3
Group 11:
Fucdil Maltose GIC(CY~-~)GICCX~-O-PAP Lactose Gal( p1-4)Glcp 1-0-PAP Lacto-N-tetraose (LNT) Gal( p l-3)GlcNAc( p 1-3)Gal(pl-4)Glc-APD Lacto-N-neotetraose (LNnT) Gal( pl-4)GlcNAc( pl-3)Gal(p1-4)Glc-APD Lacto-N-hexaose (LND)
Gal(~1-4)GlcNAc(pl-6)[Gal(pl-4)GlcNAc(pl-3)]Gal(~1-4)Glc-APD
Group 111:
Lacto-N-neohexaose (LNnD) Gal( p 1-4)GlcNAc(p 1-6)[GaI(pl-3)GlcNAc(pl-3)]Gal( p1-4)Glc-APD Melibiose Gal(a1-6)Glcpl-0-PAP Cellobiose Glc(P 1-4)Glcpl-O-PAP A-trisaccharide
GalNAc(al-3)[Fuc(al-2)]Galpl-O-APE B-trisaccharide Gal(a 1-3)[ Fuc(al-2)]Galpl-O-APE A-tetrasaccharide GalNAc(al-3)[Fuc(al-2)jGal(p1-4)Glc-APD A-heptasaccharide
GalNAc(al-3)[Fuc(a1-2)]Gal( pl-3)[Fuc(a1-4)]GlcNAc(pl-3)Gal(p1-4)GlcAPD H type 2
Fuc(al-2)Gal(pl-4)GlcNAcpl-O-APE Gangliote traose
Gal(pl-3)GalNAc(pl-4)Gal(pl-4)Glc-APD Group IV:
T-antigen Gal@ 1-3)GalNAcal-0-APE 3' Sialyllactose NeuSAc(a2-3)Gal( p1-4)Glc-APD 6' Sialyllactose
NeuSAc(a2-6)Gal(pl-4)Glc-APD LS-tetrasaccharide a NeuSAc(a2-3)Gal( p l-3)GlcNAc( p 1-3)Gal(pl-4)Glc-APD LS-tetrasaccharide b Gal( pl-3)[ (Neu5Ac(a2-6)]GlcNAc(pl-3)Gal(pl-4)Glc-API~ LS-tetrasacharide c NeuSAc(a2-6)Gal( p1-4)GlcNAc( p 1-3)Gal( p1-4)Glc-APD Disialylated lacto-N-tetraose Neu5Ac(a2-3)Gal(pl-3)[NeuSAc(ol2-6)]GlcNAc(~l-3)Gal( p1-4)Glc-APD Sialylated lacto-N-fucopentaose I1 (SLNF 11) NeuSAc(ol2-3)Gal(p 1-3)[Fuc(al-4)]GlcNAc(p1-3)Gal( pl-4)Glc-APD Sialylated lacto-N-fucopentaose I11 (SLNF 111) NeuSAc(a2-3)Gal( pl-4)[Fuc(al-3)]GlcNAc(p1-3)Gal( pl-4)Glc-APD
REACTION OF MLuCl ANTIBODY WITH THE LEY DETERMINANT
227
TABLE I (continued)
Group V: Biantennary octasaccharide Gal( p 1-4)GlcNAc(pl-2)Man(a1-6)[Gal( p 1-4)GlcNAc(p1-2)Man(a1-3)] Man@ 1-4)GlcNAc-APD Globotriose Gal(a1-4)Gal( p1-4)Glc-AF'D Globotetraose GalNAc( p 1-3)Gal(al-4)Gal( p1-4)Glc-APD Man2GlcNAc Man(al-3)Man(PI-4)GlcNAc-APD Man3GLcNAc Man(a1-2)Man(a1-3)Man( p1-4)GlcNAc-APD Chitotriose GIcNAc(~~-~)GIcNAc( pl-4)GkNAc-APD G r o w VI: Bovine serum albumin (background control)
Abbreviations:Fuc, fucose; Gal, galactose; Glc NAc, N acetylglucosamine; Glc, glucose; Gal NAc, N acetyl galactosamine; Neu 5Ac, neuraminic acid; Man, mannose. Immunostaining of high-performance thin-layer chromatograms of glycolipids Glycolipid extracts from cell lines and from normal and tumor specimens were obtained by THF extraction (Leoni et al., 1986; Magnani et al., 1987). The total glycolipid extracts and the purified Le-active ceramides (2 pg) were dried by evaporation under nitrogen stream, then dissolved in methanol and applied to silica HPTLC plates. After chromatography in chloroform/methanol/0.25% KCI (5:4:1) the plates were soaked for 1 min in 0.1% of poly(isobutylmethacry1ate) beads dissolved in hexane and immunostained by '2SI-labelledMLuCl (10' cpm/ml) for 60 min at room temperature. After extensive washing with phosphate buffer, the plates were submitted to autoradiography (Leoni et al., 1986). Inhibition of MLuCl-binding by purified oligosaccharides The effect of purified oligosaccharides on the binding of MLuCl to a total glycolipid extract of A431 cells was tested by solid-phase radioimmunoassay carried out on polyvinylchloride microtiter plates coated with the extract (Leoni et al., 1988). The amount of A431 extract was determined by a previous titration with '251-labelled MLuC1. Briefly, serial dilutions of purified oligosaccharides were incubated with 1251-labelledMLuCl (5 x lo5 cpm/well) for 1 hr at room temperature. The mixture was then transferred to glycolipidcoated plates and incubated for 2 hr. After washing with phosphate buffer the wells were cut from the plates and the bound radioactivity was determined with a gamma counter. The percentage of MLuCl binding-inhibition exerted by each oligosaccharide dilution was calculated. Direct binding of MLuCl to neoglycoproteins The binding of MLuCl to 34 different neoglycoproteins was tested by a conventional ELISA. The carbohydrate structures on the neoglycoproteins are shown in Table I. To simplify the screening procedure, the 34 different neoglycoproteins were divided into 5 groups. Wells coated with a balanced mixture of glycoproteins from each group (total amount 100 pgiwell) were used for initial screening. Each member of any positive group was then individually assayed at a concentration of 20 pg/well. Immunohistochemistly Surgical specimens were obtained from cancer patients. Pathological or normal adjacent tissues were freed of necrotic areas as far as possible and frozen in nitrogen. MLuCl reactivity on tumor-cell lines and on tissues sections was evaluated by IF and/or IPX (Tagliabue et al., 1985; Agresti et al., 1988).
FIGURE1 - Binding reactivity of '2sI-labelled MLuCl on: (a) materials separated by SDS-PAGE and blotted on nitrocellulose after solubilization from MCF7 (l), A431 (2) and SW626 (3) tumor-cell lines; (b) HPTLC-separated total glycolipid extracts from N592 (4), MCF7 ( 5 ) and A431 (6) tumor cell lines. Nitrocellulose paper and HPTLC plates were autoradiographed for 20 hr at -70°C in the presence of an intensifying screen. Molecular weight markers for SDS-PAGE are indicated on the right of panel (a)' RESULTS
Characterization of MLuCl antigenic molecules Single-cell suspensions from various human tumor-cell lines were evaluated by I F to determine the best antigenic source for the biochemical characterization of the antigen. MLuCl reacted with 11 out of 15 cell lines of epithelial origin, whereas none of the 4 cell lines of non-epithelial origin (melanoma, sarcoma and lymphoma) were reactive (data not shown). Three epithelial cell lines, i.e. MCF7 (breast carcinoma), A431 (vulvar carcinoma) and SW626 (ovarian carcinoma) were used in the following biochemical analysis. Solubilized extracts were analyzed by immunoblotting with MLuCl and the relevant results are reported in Figure la. MLuCl strongly labelled materials of heterogeneous M, (ranging approximately from less than 14,000 to over 200,000), with a unique staining pattern for each sample. On the basis of the apparent M, and broad migration, the MLuC1-positive molecules, which excccded 14,000, were considered to be glycoproteins. The
228
LEON1 b7 AL.
samples from A431 and SW626 showed a well-defined reactive band which migrates with the ion boundary. According to our previous data (Leoni et al., 1986), this band strongly suggests the presence of antigenic glycolipids. Total glycolipid extracts, obtained from 2 of the previous lines and from the N592 cell line (SCLC), were separated by HPTLC and immunostaincd by MLuC1. As depicted in Figure lb, several immunoreactive bands were detected on all 3 extracts, including that from the MCF-7 linc (lane 5), thus suggesting that the expression of the MLuCl -positive glycolipids was below the threshold value for detection by immunoblotting.
Characterization of the carbohydrate epitope recognized by MLuC1 The epitope recognized by MLuCl was characterized by 3 different independent experimental approaches. In the first, titratcd doses of purified oligosaccharides from human milk were used to inhibit thc binding of '2'I-labelled MLuCl to glycolipids extracted from the A431 tumor cell linc. As shown in Figure 2, only lacto-difucotetraose (LDFT, Ley-like hapten) was able to totally inhibit the binding and gave an IC,, of about 800 pM. LNF I1 (Lea-hapten) exerted only a slight inhibition (around 25%) at the maximum concentration tested, whereas LNF 111 (Le"-hapten), LND I (Le"-hapten) and lactose were completely ineffective. For the second experimental approach, direct binding of "'I-labelled MLuCl to purified Le-active glycolipids was carried out by immunostaining HPTL chromatograms. As shown in Figure 2 (insert), the Ley-ceramide was strongly stained, whereas the Le", Le" and Le"-ccramides wcrc completely negative. For the final approach, MLuCl was assayed by ELISA for direct binding to a panel of 34 different neoglycoproteins
FIGURE 2 - Inhibition of the binding of i2sI-labelledMLuCl by oligosaccharides. Total glycolipid extract from the A431 cell line was used as antigenic source. MLuCl binding was evaluated in the presence of the same concentrations of LDFT (y-like), LNFIT (a), LNFIII (x), LNDI (b) and lactose. For clarity the experimental points are reported only for LDFT. For the structure of the oligosaccharides see Table I:
LDFT = Fuc(al-2)Gal(B i -4)Glc 3
i
Fuca 1 Insert: binding of 12iI-labelled MLuCl on HPTLC-separated purified Ley, LeX, Lea and Leh ceramides. The more slowly migrating band present in the Ley-ceramide preparation, faintly stained by MLuC1, could be attributed to the presence of glycolipids with extended saccharide moieties.
divided into 5 groups, as shown in Table 1. MLuCl only bound to group I (Fig. 3a) and when each neoglycoprotein in group I was assayed individually, only the Ley-active neoglycoprotein (LNnD I) was reactive (Fig. 3b).
Expression of the CaMLuCl -canying glycolipids on breast tissues Since both normal and cancerous breast epithelium were positive with MLuCl (Agresti et al., 1988) the expression of CaMLuCl was further analyzed on surgical specimens by IPX and by immunoblotting, and the relevant results are reported in Table 11. By IPX normal ductal epithelial cells from mammary glands (either resting or lactating) were diffusely positive at both cytoplasmic and membrane levels in more than 60% of the tested cases. With rcgard to malignant breast tumors, about 75% of the tested cases showed either focal membrane or diffused membrane and cytoplasmic staining. By immunoblotting, a limited number of samples from rmting or hormone-stimulated normal tissues were analyzed. Out of the 4 normal and the 5 benign samples tested, 3 normal and 2 benign samples showed immunostained glycoproteins of different M, but in no case were glycolipids identified. As regards the 30 samples from carcinomas, 11 MLuCl-positive cases were found and the majority of the reactive molecules were glycoproteins. However, in 2 cases a glycolipid band was also clearly detectable. Expression of the CaMLuCI-canying glycolipids on lung tissues Since the antibody MLuCl appears to be of clinical interest in the case of SCLC (MCnard et al., 1988),we further evaluated its reactivity on normal and tumorous lung tissues. As reported in Table 111, a strong expression of CaMLuCl was observed on all the cell lines derived from SCLC, whereas the 3 lung cancer lines of different histotypes were found to be negative. At variance with the cell linc data, the majority of the specimens taken from the lung tumors were positive regardless of their histotypes. The staining pattern was, however, different, being homogeneous on SCLC and adenocarcinomas and hctcrogeneous on squamous carcinomas. Moreover, as previously reported (Agresti et al., 1988), a focal reactivity was observed on normal bronco-alveolar glands of about half of the normal lung specimens tested. Analysis of the glycoconjugates responsible for the MLuCl reactivity was performed by 2 different approaches. Solubilized extracts from 3 SCLC cell lines and a few available specimens of normal and tumorous lung tissues were analyzed by immunoblotting with MLuC1. As shown in Figure 4, on all but one (lane c ) of the SCLC samples, either from cell lines or from surgical specimens, MLuCl seemed to preferentially or exclusively identify glycolipid molecule(s), whereas on one squamous (lane g) and one adeno (lane h) carcinoma sample a faint staining was observed only on glycoproteins of different M,. According to the focal reactivity on normal tissues and to the low sensitivity of the method, no staining was detected on the 2 normal lung specimens tested (data not shown). According to cell-line analysis (see Fig. 1), the immunostaining of HPTLC-separated glycolipid extracts appeared to be more sensitive than immunoblotting in detecting the presence of MLuCl -reactive glycolipids. Therefore, in a selected series of individual patients, the expression of CaMLuCl was evaluated on the same surgical specimens by I F on frozen sections and by immunostaining on the total glycolipid extracts. As reported in Table IV, MLuCl weakly stained by IF about 10% of cells in 2 out of 4 normal lung specimens, whereas no reactivity was seen with the 4 glycolipid extracts. Similarly, all the samples except one from the 7 adeno and squamous carcinomas tested were strongly positive by IF, whereas in only 1 out of the 7 samples were MLuCl-reactive glycolipids detected.
229
REACTION OF MLuCl ANTIBODY WITH THE LEY DETERMINANT o.8 Absorbance at 650 nm -
~
_
_
_
~
~
7 _i _
1
Absorbance at 650 nm
E 0.6
0.6
d.
0.4
0.2
0 I
0.4
0.2
V
IV
111
II
0.6
BSA
0 LNnDl
Neoglycoprotein group
LNFl
LNFll
LNFlll
LNDl
Maltose
BSA
FIGURE 3 - MLuCl binding by ELISA to: (a) neoglycoprotein groups; (b) each neoglycoprotein of group I. For the structure of the oligosaccharides see Table I. TABLE I1 - MLuCl BINDING REACTIVITY ON NORMAL AND TUMOR BREAST TISSUES AS DETERMINED BY IMMUNOHISTOCHEMISTRY AND IMMUNOBLOlTING
Breast tissues
Normal Benigntumor Carcboma3
Number of surgical specimens
11
5 30
Number positiveinumber tested Immunohistochemistm' F
D
Oil1 7/11 ND5 ND 11/29 11/29
lmmunoblotting* Glycoproteins Glycolipids
314 215 11/30
014
015 24/30
'On frozen section by IPX. Intensity and percentage of positivity expressed as: F = focally positive on less than 30% of cells; D = diffusely positive on more than 70% of cells.-zOn tissue extracts. Immunostained molecules were operationally defined as: glycoproteins when separated in the separation el and glycolipids when migrated with the boundary of ions.%xAtding primary and meta~tasis.-~Thepositive cases also expressed glycoproteic molec~les.-~ND, not determined.
In contrast, in the case of SCLC specimens, good agreement could be observed between the results of the 2 tests (in 3 out of the positive cases by IF the glycolipid antigen was identified). In all the positive cases the HPTLC staining pattern was very similar to that shown in Figure l b for the cell lines. DISCUSSION
The epitope recognized by the MLuCl MAb was analyzed by 3 different independent immunochemical methods. All of the results demonstrated that MLuCl binds specifically only to the Ley-active tetrasaccharide structure. The MAb therefore seems to differ from some of the already described anti-Ley antibodies which showed a cross-reactivity with other similar structures (Fukushi et al., 1984; Gooi et al., 1985). The wide binding analysis performed on different oligosaccharides enables us to define the saccharide sequence recognized by the MAb. The results indicate that the N-acetyl group of N-acetylglucosamine is not required for MAb binding; in fact, LDFT (see legend of Fig. 3 for the structure), which contains glucose as the reducing sugar, is also able to inhibit MLuCl binding. Therefore, the proposed epitope is a type-2 chain core with 2 fucosyl substitutions. The pattern of reactivity of MLuCl with various human carcinomas and particularly with SCLC (Agresti et al., 1988) is in agreement with some (Sakamoto et al., 1986; Hirohashi et al., 1988) but not all (Zenita et al., 1988) of the reports concerning anti-Ley MAbs. This discrepancy may be due to slight variations in the epitope or in the binding affinities of the various antibodies.
TABLE 111 MLuCl IMMUNOHISTOCHEMICAL REACTIVITY ON CELL LINES FROM LUNG TUMORS AND ON NORMAL AND TUMOROUS LUNG TISSUES ~
Material Cell lines
SCLC NCI-H69 NCI-H128 N592 POVD Adenocarcinoma CaLu3 AS49 Squamous carcinoma SKMESl
MLuCl reactivitv Intensitv'
++ ++ ++ ++
% of positive cells
95 90
90 95 -
-
-
Surgical specimens
Normal SCLC Adenocarcinoma Sauamous carcinoma
316 45/52 17/21 18124
'On live cells by IF + + = homogeneous staining.-20n frozen sections by IF. In brackets, the intensity and % of positivity expressed as: F = focally positive on less than 30% of cells; H = heterogeneously positive (from 30 to 70% of cells); D = diffusely positive on 70-100% of cells. The difference in the fine specificity of the antibodies could probably also account for the different pattern of reactivity exhibited by MLuCl on A431 compared to the previously reported data (Basu et al., 1987). Analysis of the glycoconjugates which carry the Ley-related epitope recognized by MLuCl was mainly performed by immunoblotting since this procedure was previously found to be suitable for simultaneous identification of the presence of both glycoproteins and glycolipids (Leoni et al., 1986). However, its exploitation was limited by the level of sensitivity which, if compared to immunohistochemistry, was quite low. Therefore, in the case of SCLC, immunostaining of HPTLCseparated glycolipids was also applied. The analysis, although performed in a restricted number of cases and within the limits of the applied methodologies, suggested that SCLC, as compared to other lung and breast tumors and their normal corresponding tissues, preferentially exhibited glycolipid structures. This observed overexpression of Ley glycolipids on SCLC could account for the apparent discrepancy between the high frequency of increased Ley level seen in the serum of patients
230
LEON1 E T A L . TABLE IV
MLuCl REAC IIVITY ON SELECTED CASES OF NORMAL AND TUMOROUS LUNG TISSUES AS DETERMINED BY IMMUNOFIISTOCFIEMISTRY AND IMMUNOSTAINING ~
Number posilivelnumber tested
Lung tissues
lmmunohistochemistry
HPTLC immunoreaction*
Normal Tumor adenocarcinoma squamous carcinoma SCLC
214 (F)
014
:;: {El
113 014 314
414 (D)
'On frozen sections by IF. In brackets, the intensity and % of positivity expressed as: F = focally positive on less than 30% of cells; H = heterogeneously positive (from 30 to 70% of cells); D = diffusely positive on 70-100% of cells.-20n total glycolipid extract.
FIGURE 4 - Binding reactivity of 'Z'I-labelled MLuCl on material separated by SDS-PAGE and blotted on nitrocellulose after solubilization from: H69 (a), N592 (b), POVD (c) SCLC cell lines and from surgical specimens of 3 SCLC (d, e and f), 1 squamous (8) and 1 adeno (h) carcinoma of the lung. The paper was autoradiographed for 168 hr at -70°C in thc presence of an intensifying screen. Molecular weight markers are indicated on the right. with different carcinomas (Kannagi et al., 1986), and the low frequency observed in lung-cancer patients. Since B G A a r e mainly found in the serum on large mucin-like molecules (Magnani et al., 1983), it is possible that the preferential expression on glycolipids in SCLC prevents their shedding. In the case of SCLC the overcxprcssion of other glycolipids such as the hydroxy fucosyl-GM1 (Nilsson et al., 1986), GD2 (Cheresh et al., 1986) and the glycolipid recognized by the antibody SMI (Bernal et al., 1988) has already been described. T h e genetic alterations which lead to transformation are gencrally accompanied by modulation of the biosynthesis of
cell-surfacc glycoconjugates. As a consequence, the aberrant expression of glycolipid structures on tumor cells has been reported for several cancers (Hakomori, 1985). T h e glycolipids which are neosynthetized or overexpresscd by tumor cells may play a role in several processes such as cell attachment t o the matrix (Hakomori, 1981), growth regulation and response t o growth factors (Bremer et al., 1984). Thcrefore, the MLuCl M A b may b e particularly useful in the case of SCLC, where glycolipids carrying Ley are extremely abundant, to examine the mechanism of aberrant glycosylation in tumor cells and its effect o n the functions of cell-surface glycoconjugates. ACKNOWLEDGEMENTS
This work was partially supported by a grant from the Associazione per la Ricerca sul Cancro. W e thank Ms. G. Ferrami, Ms. C. Andrews and Ms. Seiler for excellent technical assistance, Ms. L. Mameli and Ms. M. Hatton for manuscript prcparation and Mr. M. Azzini for the photographic reproductions.
REFERENCES
Cell-surface antigens of human lung tumors detected by mouse AGRESTI,R., ALZANI,R., ANDREOLA, S., BEDINI,V., G I A N ~S., , MENARD, S., RILKE,F. and COLNAGHI, M.I., Histopathological charac- monoclonal antibodies: definition of blood-group and non-blood-groupterization of a novel monoclonal antibody, MLuCl, reacting with lung related antigenic systems. Int. J. Cancer, 46,1007-1013 (1990). carcinomas. Tumor;, 74,401-410 (1988). FEIZI,T., Demonstration by monoclonal antibodies that carbohydrate BASU,A,, MUKTHY, U., RODECK,U., HERLYN,M., MATTES, L. and structures of glycoproteins and glycolipids are onco-developmental DAS, M., Presence of tumor-associated antigens in epidermal growth antigens. Nature (Lond.), 314,53-57 (1985). factor receptors from different human carcinomas. Cancer Res., 47, FEIZI,T. and CHILDS,R.A., Carbohydrates as antigenic determinants 2531-2536 (1987). of glycoproteins. Biochem. J., 245,l-I1 (1987). BERNAL,S.D., CUALING, H.M., ELIAS,A D . , SCHWAKWNG, G. and FUKUSHI, Y . ,HAKOMOKI, S., NUDELMAN, E. and COCHRAN, N., Novel MCCLUER,R.M., Classification of small cell carcinoma antigens by fucolipids accumulating in human adenocarcinoma. 11. Selective isolatheir association with subcellular components. Lung Cancer, 4,99-102 tion of hybridoma antibodies that differentially recognize mono-, di-. (1988). and trifucosylated type 2 chain. J. bid. Chem., 259,4681-4685 (1984). BLASZCZYK, M., HANSSON, G.C., KARLSSON, K.A., LARSON, G., STROM- GOOI,H.C., PICAKD, J.K., H O U N S ~ LE.F., L , GREGOKIOU, M., R ~ E s , BERG, N., THURIN, J., MEENHAKD, H., STEPLEWSKI, Z. and KOPKOWSKI, A.R. and FEIZI,T., Monoclonal antibody (EGFIG49) reactive with the H., Lewis blood group antigen defined by monoclonal anti-colon epidermal growth factor receptor of A431 cells recognizes the blood carcinoma antibodies. Arch. hiochem. Riophyr., 233,161-168 (1984). group Aleh and Ale' structures. Mol. Immunol., 22,689-693 (1985). BREMER,E.G., HAKOMORI, S.I., BOWEN-POPE, D.F., RAINES,E. and HAKOMORI, S., Glycolipids in cellular interaction, differentiation, and Ross, R., Ganglioside-mediated modulation of cell growth, growth factor binding, and receptor phosphorylation. J. biol. Chem., 259, oncogenesis. Ann. Rev.Biochem., 50,733-764 (1 981). HAKOMORI, S., Aberrant glycosylation in cancer cell membranes as 6818-6825 (1984). on glycolipids: overview and perspectives. Cancer Res., 45, CHEKESH.D.A., R O S E N B ~ KJ., G ,MUJOO,K., HIRSCHOWI-IL., L. and focused REWELD,R.A., Biosynthesis and expression of the disialoganglioside 2405-2414 (1985). S., Aberrant glycosylation in tumors and tumor-associated GD2, a relevant target antigen on small cell lung carcinoma for HAKOMORI, carbohydrate antigens.Advanc. CuncerRes., 52,257-331 (1989). monoclonal antibody-mediated cytolysis. Cancer Rex, 46, 5112-5118 (1986). HAKOMORI, S., Biochemical basis of tumor-associated carbohydrate CLAUSEN, H. and HAKOMORI, S., ABH and related histo-blood-group antigens. Immunol. Allergy Clin. N. Amer., 4,781-802 (1990). antigens: immunochemical differences in carrier isotypes and their HIKOHASHI, S., KUNII,T., HIZANO,T. and SHIMOSAI'O, Y., 145-kDa distribution. Vox Sang., 56, 1-20 (1989). cells membrane antigen on SCLC as a common target for several EY.P.L., PROWSE,S.L. and JENKIN, C.R., Isolation of pure IgG,, IgG, monoclonal antibodies raised against SCLCs. Lung Cancer, 4,103-104 and IgGZ, immunoglobulins from mouse serum using prutein (1988). A-Sepharose. Inimunochemisty, 15.429-436 (1978). KANNAGI, R., FLJKUSHI. Y., TACHIKAWA, T., NODA, A,, SHIN,S., Y., INAMOTO, T., HAKOMORI, S. K., HIKAIW.4, N., FUKUDA, FEICK~RT. H.J., ANGLR,B.R., CORDON-CAKDO. C. and LLOYO,K., SHIGETA,
REACTION OF MLuCl ANTIBODY WITH THE LEY DETERMINANT
and IMURA,H., Quantitative and qualitative characterization of human cancer-associated serum glycoprotein antigens expressing fucosyl or sialyl-fucosyl type 2 chain polylactosamine. Cancer Res., 46, 2619-2626 (1986). LAEMMLI, U.K., Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (Lond.), 227,680-685 (1970). LEONI,F., MAGNANI, J.L., Mion.~,S., CANEVARI, S., PASQUALI, M., S. and COLNAGHI, M.I., The antitumor monoclonal antibody SONNINO, MOv2 recognizes the Lewis A hapten. Hybridoma, 7,129-139 (1988). LEONI,F., MIOITI,S., CANEVARI, S., SONNINO, S., RIPAMONTI, M. and COLNAGHI,M.I., Carbohydrate epitope defined by an antitumor monoclonal antibody detected on glyco roteins and a glycolipid by immunoblotting. Hybridoma, 5,289-296 8986). LOWRY,O.H., ROSEBROUGH, N.J., FARR,A.L. and RANDALL, R.J., Protein measurement with the Fohn phenol reagent. J. bid. Chem., 193,265-275 (1951). MAGNANI, J.L., Mouse and rat monoclonal antibodies directed against carbohydrates. Meth. EnzymoL, 138,484491 (1987). MAGNANI, J.L., SPITALNIK, S.L. and GINSBURG, V., Antibodies against cell surface carbohydrates: determination of antigen structure. Meth. Enzymol., 138,195-207 (1987). MAGNANI, J.L., STEPLEWSKI, Z., KOPROWSKI, H. and GINSBURG, V., Identification of the gastrointestinal and pancreatic cancer-associated antigen detected by monoclonal antibody 19-9 in the sera of patients as a much. Cancer Res., 43,5489-5492 (1983). M~NARD S.,, PORRO,G., GIAN‘L, S., BEDINI, A,, BIDOLI,P., SANTORO, A., BREGNI,M., GIANNI,A.M., PiLoi-rI, S., RILE, F., SORESI,S., FASOLATO, S. and COLNAGHI, M.I., Detection of SCLC metastatic cells
23 1
in bone marrow by means of the monoclonal antibody MLuCl. Lung Cancer, 4,73-75 (1988). MIOTTI,S., LEONI,F., CANEVARI, S., SONNINO, S. and COLNAGHI, M.I., Immunoblotting detection of carbohydrate epitopes in glycolipids and glycoproteins of tumoral origin. In: H.F. Oettgen (ed.), Gangliosides and cancer, pp. 168-176, VCH Publ., Weinheim (Germany) (1989). NILSSON, O., BREZICKA, F.T., HOLMGREN, J., SORENSON, S., SVENNERF. and LINDHOLM, L., Detection of a ganglioside HOLM, L., YNGVASON, antigen associated with small cell lung carcinomas using monoclonal antibodies directed against fucosyl-GM1. Cancer Rex, 46, 1403-1407 (1986). J.L., CUTTITTA,F., FEDORKO, J., CARNEY, ROSEN,S.T., MULSHINE, D.N., GAZDAR, A.F. and MINNA,J.D., Analysis of human small cell lung cancer differentiation antigens using a panel of monoclonal antibodies. Cancer Res., 44,2052-2061 (1984). J., FURUKAWA, K., CORDON-CARDO, C., YIN, B.W.T., SAKAMOTO, RETTIG,W.J., OETTGEN, H.F., OLD,L.J. and LLOYD,K.O., Expression of Lewis”, Lewisb, X and Y blood group antigens in human colonic tumors and normal tissue and in human-derived cell lines. CancerRes., 46,1553-1561 (1986). E., MENARD,S., DELLATORRE,G., BARRANTI, P., TAGLIABUE, MARIANI-COSTANTINI, R., PORRO,G. and COLNAGHI, M.I., Generation of monoclonal antibodies reacting with human epithelial ovarian cancer. Cancer Res., 45,379-385 (1985). A., SHIGETA, K., HIGUCHI, K., ZENITA,K., KIRIHATA, Y., KITAHARA, HIRASHIMA, K.;MURAKI,MIYAKE, M., TAKEDA, T. and KA”AGI, R., Fucosylated type-2 polylactosamine antigens in human lung cancer. Int. J. Cancer, 41,344-349 (1988).
