Cell surface glycosylation changes accompanying immortalization and transformation of normal human mammary epithelial cells

Cancer Letters, 57 (1991) 27-36 Elsevier Scientific Publishers Ireland

27 Ltd.

Cell surface glycosylation changes immortalization and transformation mammary epithelial cells

accompanying of normal human

J.W. Raka’*, F. Basolob’* *, J.W. Elliottb, J. Russob and F.R. Miller” Departments 48201

of

“Immunology

and *Pathology,

Michigan

Cancer

Foundation,

110 E. Warren Avenue,

Detroit,

Ml

(U.S.A.)

(Received 23 November 1990) (Revision received 18 January 1991) (Accepted 21 January 1991)

Summary Fluorescent lectin binding to cell surfaces was quantitatively analysed by flow cytometry on mortal human breast epithelial cells cell line immortalized MCF-lOM, the MCF-1OA derived from MCF-10M and sublines of MCF-1OA transfected with the neomycin resistance gene (MCF-1 OAneo), the c-Ha-ras protooncogene (MCF-lOAneoN), or transfected and transformed with the c-Ha-ras activated oncogene {MCF-1OAneoT). lmmortal MCF-1 OA cells bound lo-fold more peanut agglutinin (PNA) and soy bean agglutinin (SBA) than did MCF-1OM cells. Transformed MCF-1 OAneoT cells bound approximately ten times more PNA than did non-transformed protooncogene with transfected cells (MCF-lOAneoN). Treatment of the transfectants with neuraminidase abrogated the differences in PNA-binding and reduced the differences in SBA binding. SDS-PAGE separation of PNA binding glycoproteins Correspondence

Michigan Cancer

to: F.R.

Milk, Department Foundation, 110 E. Warren

of Immunology, Avenue, Detroit,

MI 48201, U.S.A. ‘Fellow from Fulbright - Hays Scholar Exchange Program. ‘Fellow from ‘Associazione ltaliana per la Ricerca sul Cancro’. l

0304-3835/91/$03.50 Published and Printed

0 1991 El sevier Scientific in Ireland

Publishers

revealed different patterns for all MCF-1OA derived sublines. Keywords: glycosylation; epithelium; immortalization;

human mammary transformation

Introduction Surface carbohydrate structures appear to play a role in cell recognition and regulatory process [33]. There is a growing body of evidence that rearrangement of these structures malignant during transformation occurs [13,35,47] and participates in expression of the malignant phenotype [6,46]. Our unique model system comprises cells isolated from normal human breast epithelium (MCF- 10M) and its spontaneously immortalized descendent (MCF-10A line) [36,39] as well as MCF-10A sublines transfected with the neomycin (neo) resistance gene (MCF-lOAneo line), the insert containing the neo gene plus c-Ha-ras protooncogene (MCF- loAneoN), or transformed with activated c-Haras oncogene (MCF-1OAneoT) [5]. Our purpose was to study lectin binding to breast epithelial cells sharing a common origin but exIreland Ltd.

hibiting progressively aberrant in vitro growth properties. The binding of fluorescent lectins to the surface of viable cells, quantitated by flow cytometry, was used for this purpose [24]. Reagqnts were chosen on the basis of their ability, to recognize putative mammary markers (PNA, GSIB4) differentiation [22,28,29] an d markers of malignant transformation of mammary cells (SBA, DBA, ConA) [6,301. Materials and Methods Cells Cells were grown in DMEM F12 media (containing 1.05 mM calcium, 5% horse serum, 10 pg/ml insulin, 100 U/ml penicillin, 10 mg/ml streptomycin, 2.5 pg/ml amphotericin B, 0.5 pg/ml hydrocortisone, 0.1 pg/ml cholera toxin, 20 ng/ml EGF) . MCF-10M cells were derived from a subcutaneous mastectomy specimen and grown in culture in low calcium conditions (0.04 mM) [36]. From this culture, which senesces after a few passages in physiological concentrations of calcium, arose the spontaneously immortalized MCF-1OA line which does not senesce [36]. The line was transfected with control inserts containing the resistance gene (neo), with or neomycin without c-Ha-ras protooncogene (MCF- loAneo and MCF- lOAneoN, respectively) [5]. Both sublines display immortal non-transformed phenotypes in vitro. The MCF-lOAneoT subline arose from transfection with the insert containing activated (codon 12 mutaand displays c-Ha-ras oncogene tion) characteristics of malignant transformation such as EGF independence and lack of contact inhibition of growth. The MCF-1OAneoT cells, which express the activated ras gene, are able to form colonies in soft agar [5]. Flow cytometry Cells growing in confluent monolayer in T 75 Corning tissue culture flasks were harvested by brief treatment with trypsin-EDTA solution. Then DME media containing 5% CS, 5% FCS was added immediately. A single cell suspen-

sion was prepared by dissociation of remaining cell aggregates by pipetting. Cells were washed once in DME 5% CS, 5% FCS and incubated for l-2 h on ice. Samples containing lo6 cells each were then washed in PBS containing 1% CS and 0.1% sodium azide. The lectins labelled with FITC were reconstituted in distilled water, aliquoted, and frozen. Aliquots were dissolved (1: lo- 1:50) just before use in PBS 1% CS, 0.1% sodium azide or the same PBS containing 0.2 M of the appropriate blocking sugar (specificity control). All the lectins and blocking sugars were purchased from Sigma (St. Louis, MO). The cells were pelleted, and incubated with 75 PI/sample of intact or sugarinactivated lectin in a concentration of lo-50 pg/ml (usually 25 pg/ml) for 40 min on ice. Then samples were washed, fixed in 1% paraformaldehyde for 20 min, and subjected to analysis by FACStar flow cytometer under 420 nm wave length of Innova 90 Argon laser light (Becton Dickinson, Mountain View, CA). The cell viability was confirmed and compensated by gating of the scatter plot. All the cell populations were analyzed simultaneously. One set of data was generated for the whole population of each cell type, another one for the cells of approximately the same size selected from each population by a common narrow, vertical, rectangular gate. Analysis of the results was based on comparison of histograms (whole populations) or numerical data (size compensated data). The parameter used for expression of numerical differences was mean specific fluorescence, calculated by subtraction of mean fluorescence intensity of cells stained with inactivated lectin from mean fluorescence intensity of cells stained with intact lectin. The percentage of positive cells was obtained by gating the histograms beyond the range of controls. All statistical parameters were calculated automatically by the cytometer computer. Significance of differences was confirmed at the confidence level of P < 0.001 according to Kolmogorov-Smirnov statistics. Neuraminidase treatment Cells were suspended

of viable cells in 0.1 U/ml neura-

29 no neuraminidase

minidase type V, from Clostridium perfringens (Sigma, St. Louis, MO) diluted in Waymouth’s media supplemented with 2% CS, or with media alone (control), and incubated for 15 min at 37”C, followed by several washings. Western blotting and lectin overlays Cells were grown in lOO-mm tissue culture dishes (Corning Glass Works, Corning, NY) until subconfluent, then harvested with 2 mM EDTA and a cell scraper (Corning). The cell suspension was washed with PBS, pelleted and treated with lysis buffer containing 0.5% NP-40,l mM EDTA, and 1 mM PMSF in PBS (CA*+- and Mg*+- free) for 30 min on ice. The protein equivalent of 3 x lo5 cells was subjected to SDS-PAGE electrophoresis in 8% or 10% gel under reducing conditions. The protein was transferred to nitrocellulose filters. The filters were blocked in 1% horse hemoglobin (Sigma, St. Louis, MO) overnight in the presence or absence of 1 mu/ml neuraminidase type V (Sigma). After neuraminidase was removed by washing, the filters were incubated for 1 h with 2.5 pg/ml PNA in PBS containing Ca*+ and Mg*+ ions and 0.1% BSA. Prior to l-h incubation with anti-PNA rabbit antibody (Accurate Chemicals and Scientific Corp., Westbury, NY) in dilutions of 1:400 to 1:800 (usually 1:600) the filters were washed three times in PBS containing 0.1% BSA and three times (25-20 min each) in PBS containing 5% BSA fraction V - 98% purity (Sigma). Finally, after 3 washes in PBS 5% BSA and 3 washes in quench solution containing 15% skim milk in PBS, the filters were incubated for 1 h with ‘251-labelled (Amersham donkey anti-rabbit antibody Corp., Arlington Heights, IL) in dilutions adjusted to actual radioactivity (mostly 1 PI/I ml). The blots were then washed in quench solution four times (last two washings in the presence of 0.1% Tween-20) and autoradiographed using Kodak Xomat equipment. Results The cell surface PNA, SBA and SConA bin-

MCFlOM

MCFlOA

reo CO118

neuraminidase treatment BOO ,

2 :

600 -

2

n q q

5 400 I : :

PNA+neur SBA+nwr SConA+neur

i ; MCFloM

MCFlOA

ma

n&4

WOT

cdl.

Fig. 1. The lectin binding patterns of MCF-IO cell populations. The data have been collected from several experiments (lectin concentrations 25 pg/ml), normalized and expressed as a mean specific fluorescence intensity obtained by using cell size compensated values.

ding profiles of the components of our panel of cells are summarized in Fig. 1. SconA staining intensity of t-as gene containing MCF-lOAneoN and MCF-1OAneoT was similar and twofold stronger than that of MCF-lOM, MCF-10A and MCF-1OAneo cells. Although neuraminidase treatment increased the lectin binding, the relative difference remained the same. The level of binding of Dolichos biflorus agglutin (DBA) and GSIB4 lectin to all the cell populations was very low regardless of neuraminidase treatment (data not shown). PNA recognizes carbohydrates with terminal galactose residues [15,43-451. The binding

MCF-1OM

0

,, : . .

MCF-1 OAneoN

. . . .

FL1

FL1

300

MCF-1 OAneoT

FL1

FL1

f

100

PNA, % positive cells

80

0 MCFloM

FL1

Fig. 2.

PNA binding profiles of MCF-10 cell populations. Panels a-e: population. Cells were incubated with 25 pg/ml PNA-FITC: solid lines. solution containing 0.2 M galactose (densely dotted lines). The sparsely of cells treated with neuraminidase. Panel f: Percentages of positive cells treatment with neuraminidase.

MCFl

OA

neo

WJN

fle0T

cells

The relative fluorescence intensity of the whole Control samples were incubated with the same dotted lines represent the lectin binding profiles before (solid bars) and after (cross-hatched bars)

31

profiles of PNA, in a concentration of 25 pg/ml, to the MCF-10 cell populations are shown in Fig. 2. The lectin preincubated with 0.2 M D( + )-galactose did not display any specific binding activity in a range of concentrations up to 50 pg/ml (histograms in the range of autofluorescence). The intact lectin bound to different cell populations with different efficiencies reflected by mean specific fluorescence intensities. There was heterogeneity (wide histograms) within the positive populations: MCF-IOA, and MCF-10AneoT. Neuraminidase treatment increased the binding efficiency causing the shift of the histograms as well as an increase in the percentage of positive cells up to 100%. In the absence of neuraminidase treatment, MCF-10A

NO NEURAMINIDASE

Fig. 3.

NEUNAMINIDASE

cells displayed 3- 10 times stronger fluorescence than did the MCF-1OM parental population. After desialylation the fluorescence greatly increased and the difference was reduced or even abrogated in some experiments. PNA binding to MCF-1OAneo and MCF-lOAneoN transfectants was negligible unless the cells were treated with neuraminidase. In contrast, the MCF-lOAneoT transfectant containing activated ras oncogene bound more PNA (4-100 times more, in different experiments) than either control transfectant. All three transfectants expressed similar and much increased PNA binding capacity after desialylation but the level of the lectin binding remained about twofold lower than in the case of the MCF-10A line.

TREATMENT

The SDS-PAGE separation of PNA binding glycoproteins from lysates of the MCF-1OA line and transfectants. The whole cell lysates obtained from equal cell number of each population were separated in 8% gel, transferred to nitrocellulose which had then been treated (or not) with neuraminidase. The proteins were visualized by autoradiography following overlay with PNA, anti-PNA antibody, and the secondary radiolabelled antibody.

32

The level of PNA binding per cell was slightly higher in the case of confluent MCF-IOA and MCF-lOAneoT, but not in the case of MCF-lOAneoN cells, in comparison to cells in logarithmic phase of growth (not shown). The MCF-IOA line and transfectants differed not only with respect to the total amount of PNA binding sites per cell, as shown by flow cytometry, but also in the pattern of PNA binding proteins obtained from the whole cell lysate and separated by SDS-PAGE (Fig. 3). There were only a few molecular species weakly binding PNA without desialylation. The distribution of the high molecular weight species (> 200 kDa) seemed to parallel flow cytometry data since it was found only in cells. and MCF-1OAneoT MCF- 10A Glycoprotein of 65-75 kDa was expressed only by MCF-1OA and MCF-lOAneoN cells, whereas 55-65 kDa glycoprotein was found in MCF-1OA cells and in all three transfectants. Treatment of filters with neuraminidase revealed a variety of PNA binding glycoproteins common for all the lines (between > 200 and 85 kDa). Strong bands, however, of 65-75 kDa were found in lines MCF-lOAneoN and MCF-10AneoT. The high density band of 55-65 kDa was observed only in MCF- 1OAneo and MCF- lOAneoN. SBA is specific for terminal N-Acetylgalactosamine (Tn antigen). There is a similarity in the PNA and SBA binding patterns across the MCF-10 cell family (Fig. 1). Nearly all MCF-10A cells (99.2%) bound SBA but only 12.7% of the MCF-1OM cells were SBAbinding. The difference in mean specific SBA MCF- 10M and between fluorescence MCF-10A cells was lo-fold in favor of the latter Neuraminidase treatment increased the percentage of positive MCF-1OM cells to nearly 100% and increased the lectin binding approximately five times in the case of MCF-1OA and 20 times for MCF-10M cells, reducing the difference in SBA binding between the populations to approximately 2.4-fold. MCF- 1CAneo and MCF-lOAneoN transfectants bound 3-5 times less SBA than did the MCF-10A iine. However, the MCF-1OAneoT transformant

bound 2.5 to 3 times more lectin than the MCF-lOAneoN counterpart. Neuraminidase treatment increased the SBA binding capacities approximately 3, 9, 8 and 5 times to MCF-IOA, MCF-lOAneo, MCF-lOAneoN, and MCF-1OAneoT cells, respectively, reducing the differences the lectin binding to approximately 2-fold and increasing the percentage of positive cells from 94.9%) 62.5%) 36.4% and 88.6%, respectively, to 100% in all cases. Discussion A variety of carbohydrate modifications accompanying malignant transformation in different organs have been reported [13,35,38,47]. In particular, human breast cancer cells were found to express different lectin binding patterns than do their normal counterparts [9,14,20,22,41]. However, the findings based on cytochemical or histochemical analysis display a great deal of variability. On the other hand, cultured normal breast epithelium subjected to treatment with 7,12-dimethylbenz[a]anthracene or N-methylN-nitrosourea acquires a quasi-malignant phenotype paralleled by increased ConA, SBA and DBA binding [30]. It has been documented that culturing of the breast epithelial cells as well as malignant transformation leads to expression of several oncogenes

VI. Little is known of the influence of oncogenes on expression of glycoconjugates even though some oncogene products are themselves glycoproteins [ 1,23,27,32] and glycosylafion may be important for their biological activity [23,27]. Several regulatory molecules including hormones [17], growth factors [19] and integrins [25], which can be involved in some steps of carcinogenesis, also require glycosylation for their biological function. Apparently some specific modifications of surface carbohydrates are essential for tumor initiation and progression [3,6,31,46]. In transfection experiments modification of glycosylation pattern may occur when the insert contains information for carbohydrate specific enzymes

33

[l&3]. Sugar composition can also be indirectly changed by oncogenes which do not have such an enzymatic activity [3,6,21,26,31,37]. This is probably the case in our system where sequential immortalization, transfection, and cHa-ras induced transformation are accompanied by changes in lectin binding patterns. The PNA binding profiles of mortal (MCF- lOA), trans(MCF- 10M), immortal fected (MCF-lOAneo, MCF-lOAneoN) and transformed (MCF-10AneoT) cells were different Immortalization was accompanied by an increase in the lectin binding capacity. The difference between mortal and immortal cells was not a result of cell size since mortal cells are larger but stained weaker, and because the analysis of the cells of the same size separated from both populations by gating of scatter plots still revealed stronger fluorescence of the immortalized MCF-10A line. The majority of PNA binding sites are cryptic since desialylation greatly increased the staining intensity of all the MCF-10 populations. Since neuraminidase treatment greatly reduced the difference between mortal and immortal cells, the level of sialylation seems to be the major reason for differential PNA binding by the cells. The PNA binding profile of MCF-10 cells resembles the lectin binding pattern of mammary ductal cells reported in the literature [22,28,29,42]. Further studies are necessary for explanation of the difference between mortal and immortal cells. Transfection of MCF-10A cells with plasmids containing neomycin resistance gene, c-Ha-r-as protooncogene, plus or minus resulted in profound decreases in expression of both nonsialylated and cryptic PNA binding sites. This observation is in line with Yoshikura’s finding on the role of neomycin resistance gene transfection in carbohydrate mediated phenomena is probably due to phosphorylation of those sugars by the gene specific phosphotransproduct (neomycin ferase) [48]. In comparison to control transfectants (MCF-lOAneo, MCF-lOAneoN) the PNA binding capacity of the transformed MCF-1OAneoT line was substantially elevated.

Neuraminidase treatment abrogated the difference between transformed MCF- 1OAneoT cells and control transfectants but not between immortalized MCF-1OA line and transfected sublines. This observation indicates that transfection itself is accompanied by a decrease in the amount of PNA reactive glycoconjugates whereas transformation by c-Ha-ras oncogene results in decrease in the level of sialylation or/and expression of sialylated PNA binding carbohydrates. We have found a 3-6-fold lower activity of sialyltransferase in homogenates of MCF-1OAneoT in comparison to MCF-lOAneoN cells (Woynarowska and Rak, unpublished). Separation of PNA binding glycoproteins by SDS-PAGE revealed no single alteration. The presence of high molecular weight species in MCF-10A and MCF-1OAneoT cells correspond to surface fluorescence data. There were, however, two low molecular weight bands (65-75 and 55-65 kDa) which were differentially expressed by the cell lines after neuraminidase treatment. The latter changes in PNA binding patterns may reflect the alterations of synthesis of cytoplasmatic glycoproteins since they do not correlate with flow cytometry data. The tempting candidate for further studies seem to be the high molecular weight species because of its possible relationship to PNA binding mucins [lo, 121. Although our results are consistent with previous reports [30] that transition to a transformed phenotype may be accompanied by increased SBA binding capacity (MCF-1OAneoT cells), we also observed a marked increase in SBA binding at immortalization (compare MCF-1OM and MCF-IOA). Increased concanavalin A reactivity of malignant vs. normal cells was reported for murine mammary tumors [18] as well as other tumor systems [34]. In the MCF-10 system, cells transfected with either ras protooncogene or activated ras oncogene stained stronger with SConA than the other cell populations. The Grijjonia simplicifolia I lectin apparently recognizes a-galactosides important for tumorigenicity [40], metastatic potential [4], or lec-

34

tin dependent cell mediated cytotoxicity phenomena [7,16] in a variety of murine tumor systems. GSIB4, however, bound to neither normal nor transformed MCF-10 human breast cell populations. Neither could we show any appreciable staining of the cells with Dolichos biflorus agglutinin. This is the first report showing altered glycosylation in human breast epithelium derived cells subjected to sequential immortalization and transformation with oncogene. Each of the cell populations in this system display a unique lectin binding pattern. Further studies of lectin binding may reveal additional modifications. The specific carbohydrate biological significance of differential lectin binding is to be determined. The molecular basis of this phenomenon is currently under study. In this regard glycoprotein and glycolipid profiles of MCF-10 cells are of interest, as well as activities of glycosyltransferases.

4

5

6

7

8

Acknowledgements This work was supported by UPHS grants CA38921 and CA28366. The authors would like to express their sincere thanks to Dr. Gloria Heppner and Dr. Sam Brooks for valuable critical comments on our work, as well as to Dr. Stuart Ratner and Mr. Gary Bora from the MCF FACS facility, who were extremely cooperative and provided us excellent assistance. Special thanks are owed to Dr. Avraham Raz who shared with us his knowledge, experience, and resources. We also appreciate the help of Mrs. Margaret Peterson who typed the manuscript.

9 10

11

12

References Anderson, S.J., Gonda, M.A., Rettenmier, C.W. and Sherr, C.J. (1984) Subcellular localization of glycoproteins encoded by the viral oncogene v-fms. J. Viral., 51. 730-741. Benz, C.C., Scott, G.K., Santos, G.F. and Smith, H.S. (1989) Expression of c-myc, c-Ha-rasl and c-erbB-2 proto-oncogenes in normal and malignant human breast epithelial cells. J. Natl. Cancer Inst., 81, 1704-1708. Bolscher, J.G.M., Van der Bijl, M.M.W., Neefhes. J.J., Hall, A., Smets, L.A. and Ploegh. H.L. (1988) Ras (pro-

13

14

15

to)oncogene induces N-linked carbohydrate modification: temporal relationship with induction of invasive potential. EMBO J., 7, 3361-3368. Castronovo, V., Foidart, J.M., Li Vecchi, M, Foidart, J.B., Bracke, M., Mareel, M. and Mahieu, P. (1987) Human anti-alpha-galactosyl IgG reduces the lung colonization by murine MO4 cells. Invasion Metastasis, 7, 325-345. Ciardello, F., McGeady, M.L., Kim, N., Basolo, F., Hynes, N., Langton, B.C., Yokozaki, H., Saeki, T., Elliott, J.W., Masui, H., Mendelsohn, J., Soule, H., Russo, J. and Salomon, D.S. (1990) Transforming growth factor OL expression is enhanced In human mammary epithelial cells transformed by an activated c-Ha-ras protooncogene but not by the c-ceu protooncogene, and overexpression of the transforming growth factor-o complementary DNA leads to transformation. Cell Growth Differentiation, 1, 407-420. Dennis, J.W. and Laferte, S. (1989) Oncodevelopmental expression of -GlcNAc bl-6Manblbranched asparagine linked oligosaccharides in murine tissues and human breast carcinomas. Cancer Res., 49, 945-950. Eckhard, A.E., Malone, B.N. and Goldstein, I.J. (1982) Inhibition of Ehrlich ascites tumor growth by Griffonia simplicifolia I lectin in viva. Cancer Res., 42, 2977-2979. Ernst, L.K., Rajan, V.P., Larsen, R.D., Ruff, M.M. and Lowe, J.B. (1989) Stable expression of blood group H determinants and GDP-L-fucose: beta-D-galactoside 2-alpha-L-fucosyltransferase in mouse cells after transfection with human DNA. J. Biol. Chem., 264, 3436-3447. Franklin, W.A. (1983) Tissue binding of lectins in disorders of the breast. Cancer, 51, 295-300. Gendler, S.J., Burchell, J.M.. Duhig, T., Lamport, D., White, R., Parker, M. and Taylor-Papadimitriou, J. (1987) Cloning of partial cDNA encoding differentiation and tumor-associated mucin glycoproteins expressed by human mammary epithelium. Proc. Natl. Acad. Sci. USA, 84. 6060-6064. Glick, M.C., De-Santis, R. and Santer, U.V. (1985) Glycosylation changes in membrane glycoproteins after transfection of NIH 3T3 with human tumor DNA. Prog. Clin. Biol. Res., 175, 229-237. Griffith, B., Bobrow, L.G., Harperfield. L. and Swallow, D.M. (1988) Expression of the hypervariable PUM locus in normal and malignant lung: the tumour-associated epitopes are present but masked in normal tissue. Dis. Markers, 6, 195-202. Hakomori, S-I. (1989) Aberrant glycosylation in tumors and tumor-associated carbohydrate antigens. Adv. Cancer Res.. 52, 257-331. Howard, D.R.. Ferguson, P. and Batsakin, J.G. (1981) Carcinoma-associated cytostructural antigenic alterations: detection by lectin binding. Cancer, 47, 2872-2877. Lotan, R. and Sharon, N. (1978) Peanut (An&is hypogeo) agglutinin. In: Methods in Enzymology, Vol. 50, pp. 361-367. Editor: V. Ginsburg. Academic Press, New York.

35

16

17

18

19

20

21

22

23

24

25

26

munol., 71. 202-207. Matzuk, M.M. and Boime, I. (1988) The role of the asparagine-linked oligosaccharides of the alpha subunit in the secretion and assembly of human chorionic

30

gonadotropin. J. Cell. Biol., 106, 1049-1059. Mitani, H., Murakami, A. and Tanaka, H. (1984) Differnet concanavalin A binding patterns of malignant and nonmalignant mouse mammary epithelia in monolayer

31

culture. Cancer Res., 44, 674-680. Miyazono, K. and Heldin, C-H. (1989) Role of carbohydrate structures In TGF-bl latency. Nature, 338, 158-160. Muller-Holzner, E, Marth, C, Kofler, E, Daxenbichler, G. and Hofstadter, F. (1985) Lectin binding sites in cultured human breast cancer cells. Breast Cancer Res. Treat., 5, 277-283. Nakaishi, H.. Sanai, Y ., Shibuya, M., Iwamori, M. and Nagai, Y. (1988) Neosyntheses of neolacto-cell line brought about by transfection with the v-fes oncogenecontaining Gardner-Arnstein strain feline sarcoma virus DNA. Cancer Res., 48, 1753-1758. Newman, R.A., Klein, P.J. and Rudland, P.S. (1979) Binding of peanut lectin to breast epithelium, human carcinomas, and cultured rat mammary stem cell: Use of the lectin as a marker of mammary differentiation. J. Natl. Cancer Inst., 63, 1339- 1345. Nichols, E.J., Manger, R., Hakomori, S., Herscovics. A. and Rohrschneider, I.R. (1985) Transformation by the vfms oncogene product: role of glycosylational processing and cell surface expression. Mol. Cell Blol., 5. 3467-3475. Notter, M.F.D. and Leary, J.F. (1987) Surface glycoproteins of differentiating neuroblastoma cells analysed by lectin binding and flow cytometry. Cytometry, 8, 518-525. Oz. O.K., Cambell, A. and Tao, T.-W. (1989) Reduced cell adhesion to fibronectin and laminin is associated with altered glycosylation of bl integrins in a weakly metastatic glycosylation mutant. Int. J. Cancer, 44, 343-347. Pierce, M. and Arango, J. (1986) Rouse sarcoma virus transformed baby hamster kidney cells express higher tri-and tetraantennary levels of asparagine-linked glycopeptides containing /GlcNAc-b(1,6)Mana(l,@Man/ sequences than and poly-N-acetyllactosamine J. Biol. Chem.. hamster kidney cells.

27

28

29

types in primary munohistochemical

Maddox, D.E., Goldstein, I.J. and Lobuglio, A.F. (1982) Grtffonia simplicifolia I lectin mediates macrophageinduced cytotoxicity against Ehrlich ascites tumor. Cell. Im-

Biol., 4, 1999-2009. Rudland, R.S. (1987) Stem cells and the development of mammary cancers in experimental rats and in humans. Cancer Metastasis Rev., 6, 55-83. Rudland, P.S., Hughes, C.M., ton, M. J, (1989) Characterization

33 34

35

36

Ferns, S.A. and Warburof human mammary cell

cells treated

with carcinogens

in vitro.

imIn

f 1988) breast Cancer

Res., 48, 2837-2857. Santer, U.V., De Santis, R., Hard, K.J., van Kuik, J.A., Vliegenthart. J.F., Won, B. and Glick, M.C. (1989) Nlinked oligosaccharide changes with oncogenic transformation require sialylation of multiantennae. Eur. J. Biochem., 181, 249-260. Schmidt, J.A.. Beug, H. and Hayman. M.J. (1985) Effects of inhibitors of glycoprotein processing on the synthesis and biological activity of the erb-B oncogene. EMBO J., 4. 105-112. Sharon, N. and Lis, H. (1989) Lectins as cell recognition molecules. Science, 246, 227-246. Shoham, J. and Sachs, L. (1972) Difference in the binding of fluorescent concanavalin A to the surface membrane of normal and transformed cells. Proc. Natl. Acad. Sci. U.S.A., 69, 2479-2482. Smets, L.A. and Van Beel, W.P. (1984) Carbohydrates of the tumor cell surface. Biochim. Biophys. Acta, 738. 237-249. Soule, H., Maloney, T.M., Wolman, S.R., Peterson, Jr. W.D.. Brenz, R., McGrath, C.M., Russo, J., Pauley, R.J., Jones, R.F. and Brooks, S.C. (1990) isolation and characterization of spontaneously immortalized human breast epithelial cell line, MCF-10, Cancer Res., 50,

37

38

39

40

baby 261.

1077210777. Roussel. M.F., Rettenmier, C.W., Look, A.T. and Sherr, C.J. (1984) Cell surface expression of v-fms-coded glycoproteins is required for transformation. Mol. Cell

Vitro Cell. Dev. Biol., 25, 23-36. Russo, J., Reina, D., Frederick, J. and Russo. I.H. Expression of phenotypical changes by human epithelial

32

culture: Immunofluorescent and indicators of cellular heterogeneity.

41

42

6075-6086. Spitalnik, S.L.. Spitalnik. P.F., Dubois, C., Mulshine. J., Magnani, J.L., Cuttitta, F., Civin, C.I., Minna, J.D. and Ginsburg, V. (1986) Glycolipid antigen expression in human lung cancer. Cancer Res.. 46, 4751-4755. Springer, G.F., Taylor, C.R., Howard, D.R., Tegtmeyer, H., Desai, P.R., Murthy. S.M., Felder, B and Scanlon, E.F. (1985) Tn, carcinoma-associated antigen, reacts with anti-Tn of normal human sera. Cancer, 55, 561-569. Tait, L., Soule, H. and Russo, J. (1990) Ultrastructural and cytochemical characterization of an immortalized human breast epithelial cell line, MCF-10. Cancer Res.. 50, 6087-6094. Varani, J., Lovett III, E.J., Wicha, M., Malinoff, H. and McCoy Jr., P. J (1983) Cell surface o-D-galactopyranosyl end groups: Use as markers in the isolation of murine tumor cell lmes with different cancer-causing potentials. J. Natl. Cancer inst., 71, 1281-1286. Walker, R.A. (1985) The binding of peroxidase-labelled lectins to human breast epithelium. IV. The reactivity of breast carcinomas to peanut, soy bean and Dolichos biflorus agglutinins. J. Pathol., 145, 269-277. Warburton, M. J., Ferns, S.A., Hughes, C.M. and Rudlan, P.S. (1985) Characterization of rat mammary cell types in primary culture: lectins and antisera to basement membrane and intermediate filament proteins as indicators of cellular heterogeneity. J. Cell Sci., 79, 287-304

36 43

44

45

Wolf, M.F., Koerner, U. and Schumacher, K. (1986) Specificity of reagents directed to the ThomsenFriedenreich antigen and their capacity to bind to the surface of human carcinoma cell lines. Cancer Res., 46, 1779--1782. Wolf, M.F., Schmitt, H.R. and Schumacher, K. (1989) Expression of Thomsen Friedenreich (TF) antigens on lymphocytes. I. Distribution of cryptic and exposed TF antigens on murine lymphocytes from different lymphoid organs: detection with an anti-TF monoclonal antibody and peanut agglutinin. Cell. Immunol., 121, 360-365. Wolf, M.F.. Schmitt, H.R.. and Schumacher, K. (1989) Expression of Thomsen Friendenreich (TF) antigens on lymphocytes.

11. Loss

of

cryptic

TF

antigens

during

46

47

48

mitogenic activation of human T and B lymphocytes. Cell. Immunol., 121, 366-371. Yagiel, S., Feinmesser, R., Waghorne, C., Lala, P.K., Breitman, M.L. and Dennis, J.W. (1989) Evidence that bl-6 branched Asn linked oligosaccharides on metastatic tumor cells facilitate invasion of basement membranes. Int. J. Cancer. 44, 685-690. Yogeeswaran, G. (1983) Cell surface glycolipids and glycoproteins in malignant transformation. Adv. Cancer Res., 38. 289-350. Yoshikura, H. (1989) Suppression of focus formation by bovine papilloma non-transformed phosphorylation.

virus-transformed cells by contact cells: Involvement of sugar(s) Int. J. Cancer, 44. 885-891.

with and