Different Patterns of Lectin Binding and Cell Surface Sialylation Detected on Related High- and Low-Metastatic Tumor Lines

(CANCER RESEARCH 43, 5138-5144, November 1983]

Peter Altevogt,1 Mina Fogel, Rachanee Cheingsong-Popov, Jim Dennis, Peter Robinson, and Volker Schirrmacher2

Institut fur Immunologie und Genetik, Deutsches Krebsforschungszentrum, 6900 Heidelberg, Federal Republic of Germany [P. A., M. F., R. C-P. P. R., V. S.Â¡,and The Hospital for Sick Children, Toronto, Ontario, Canada [J. D.J

ABSTRACT tation of metastatic tumor cells by heterotypic cell interactions; and (c) in the arrest in organs by specific interactions with the We have analyzed cell surface-bound carbohydrates in two target tissue. different model systems for metastasis composed of closely Such important involvements of carbohydrates in metastasis related tumor cell lines with differing metastatic potential. The formation would imply that tumor cells with different metastatic first system studied was that of the DBA/2-derived T-lymphoma potential possess distinct qualitative or quantitative differences lines (Eb/ESb) and some recently established sublines of ESb in their glycosylation of membrane glycoconjugates. A correlation with altered metastatic behavior (ESb-M and ESb-MR). The between the quantity of cell surface-exposed SA3 and metastatic second system consisted of the highly metastatic MDAY-D2 capacity was recently found by Yogeeswaran (36) for a variety cells, a wheat germ agglutinin-resistant low metastatic subline of different tumor cell lines. Of particular importance seemed to MDW40, and two metastatic revertants from the latter. The cells be SA residues in association with subterminal sugar moieties were stained with fluorescein isothiocyanate-conjugated lectins which could be oxidized by galactose oxidase and subsequently and analyzed by flow cytofluorography. All low-metastatic tumor reduced with NaBH4 (36). lines expressed receptor sites for the lectins soybean agglutinin We have further investigated this topic with cell lines of 2 (SBA) and Vicia villosa (VV). The metastatic lines had the re independent tumor model systems for metastasis. Differences in spective lectin binding sites blocked by sialic acid (SA). A good glycosylation and the role of SA were studied by using a variety correlation was found within the cell lineages Eb â€”Â»ESbâ€”Â»ESb- of FITC-conjugated lectins with different sugar specificities and M â€”ESb-MR and MDAY-D2 -Â»MDW40 -Â»MDW40M1 between flow cytofluorographic analysis. The first tumor system was the reactivity of SBA and VV and metastatic potential. DBA/2 T-lymphoma Eb and its spontaneous high-metastatic The amount of neuraminidase-accessible SA was similar for variant ESb (30). It was found that acceptor sites of SBA and all cell types (except MDW40) indicating differences in the posi VV were covered by SA on all high-metastatic ESb sublines but tioning of SA. For high-metastatic ESb cells, the Sialylation of not on parental type Eb cells or on low-metastatic variants from SBA and VV receptor sites was paralleled by a relative decrease ESb cells. Similar differences were noted in a second tumor of SA associated with receptor sites for peanut agglutinin. Low- system (9, 11, 18, 19). From these results, we suggest that the metastatic Eb cells, in contrast, had their peanut agglutinin Sialylation of particular carbohydrate residues on tumor cells receptor sites sialylated but expressed asialylated SBA and VV which are recognized by certain lectins may facilitate their met receptor sites. Eb cells were also found to have 2-fold higher astatic spread possibly by masking specific cellular adhesion activities in galactose-specific sialyltransferases. It is proposed sites. that the differences in positioning of SA on the cell surface leading to masking or unmasking of terminal sugars could influ MATERIALS AND METHODS ence the metastatic potential of tumor cells. Cells. The tumor cell Eb (Heidelberg subline of L5178Y) and its metastatic variant ESb were obtained from Professor P. Alexander INTRODUCTION (London, England). The etiology of the cell lines has been described (30). Both cell lines were passaged as ascites in syngeneic DBA/2 mice Alterations in cell surface carbohydrates have been found to (Bomholtgaard, Denmark) or grown in suspension in Roswell Park Mem severely affect the metastatic potential of experimental tumors. orial Institute tissue culture Medium 1640 supplemented with penicillin For example, tumor cell variants selected in the presence of (100 units/ml), streptomycin (100 /jg/ml), glutamine (5 HIM), 4-(2-hydrox- toxic concentrations of certain plant lectins appeared to have yethyl)-1-piperazineethanesulfonic acid (50 mM), 2-mercaptoethanol (5 x 10~5 M), and 5 or 10% fetal bovine serum. ESb-M is a plastic-adherent significantly changed metastatic behavior (11, 28, 34). Such selection procedures can induce changes in specific surface variant of ESb with highly decreased metastatic capacity (14). MDAY- carbohydrates (11, 12), glycosyltransferase activities (12, 15, D2 is a metastatic tumor cell line of DBA/2 and was established by Kerbel ef al. (18,19). A WGAR cell line MDW40, with altered tumorigenic- 23), and glycoprotein content (22). One could theoretically envis age at least 3 steps in the metastatic cascade that involve cell ity and metastatic capacity (9), and 2 metastatic and wheat germ agglutinin-sensitive revertants thereof (MDW40M1 and MDW40M4) have surface carbohydrate interactions: (a) in the release of tumor been described previously (11 ). cells from the primary tumor mass due to altered homotypic Staining of Tumor Cells with FITC-conjugated Lectins and Cyto adhesion phenomena; (b) in the mechanism of blood transpor- fluorographic Analysis. FITC-conjugated lectins (1 mg/ml) were ob tained from Medac (Hamburg, Federal Republic of Germany). Fluoro- 1To whom requests for reprints should be addressed. Recipient of a grant from Meyenburg-Stiftung. 3The abbreviations used are: SA, sialic acid; FITC, fluorescein isothiocyanate; * Supported by the Deutsche Forschungsgemeinschaft through Sonderfor- SBA, soybean agglutinin; VV, Vicia vil/osa: WGA", Wheat germ agglutinin-resistant; schungsbereich 136. HP, Helix oomatia, GalNAc, N-acetylgalactosamine; PNA, peanut agglutinin; o-Gal. Received September 2, 1982; accepted August 8, 1983. D-galactose; HBP, hepatocyte-binding protein; GlcNAc, N-acetylglucosamine.

5138 CANCER RESEARCH VOL. 43

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1983 American Association for Cancer Research. Cell Surface Glycosylation and Metastatic Potential chrome/protein ratios (495/280 nm) were between 1.3 and 1.8. Cells absent on ESb tumor cells but present on Eb cells (1). (106) were incubated with 50 n\ of 1/20 diluted FITC-lectins in the absence Staining of Eb cells with SBA, VV, and HP was only slightly or presence of 0.2 M competitive sugars on ice for 45 min. The reaction increased after neuraminidase treatment, indicating that these was carried out in 0.9% NaCI solution supplemented with 1 HIM calcium receptor sites were mostly unsialylated and accessible for the and magnesium salts. Cells were washed and analyzed by flow cytofluorography using an ORTHO 50-H cell sorter. The argon laser was operated lectins (Chart 1). at 488 nm, and 3 x 104 cells were analyzed using settings to exclude ESb cells showed weak reactivity with PNA, while Eb cells nonviable cells according to scatter characteristics. Gains were kept at were negative as judged by the histograms in Chart 1. After identical settings for each FITC-lectin analyzed. neuraminidase treatment, both cell lines were stained by PNA, Neuraminidase Treatment of Cells. Cells were incubated in Dulbec- but the ESb cells still showed stronger reactivity as seen by the co's phosphate-buffered saline at 20 to 40 x 106/ml for 45 min at 37Â°in percentage of stained cells and by the fluorescence intensity. In the presence of 2.5 units of neuraminidase of Vibrio cholerae (Behring- contrast, Ricinus communis agglutinin was able to bind to both werke, Marburg, Federal Republic of Germany) and were then washed Eb and ESb cells, possibly via unblocked terminal o-galactose 3 times and analyzed for lectin binding as described above. residues. This binding was enhanced after treatment with neur SA Determination. Fifteen to 20 x 106 cells/ml were treated with 2 units of neuraminidase for 45 to 60 min at 37Â°.Neuraminidase-releasable aminidase and most strongly for Eb cells, indicating a higher degree of o-galactose sialylation (Chart 1S). No major differences SA of the supernatants was determined by the thiobarbiturate method in lectin binding properties of Eb and ESb cells were noted for u- of Warren (35). fucose- or o-mannose-detecting lectins [see results in Chart 16 Determination of Glycosyltransferase Activities. Cell homogenates with Ulex europaeus (L-fucose) or concanavalin A (o-mannose)]. from Eb and ESb were prepared according to the method of Chatterjee ef al. (8). One g of packed cells was resuspended in 3 ml of 0.25 M The results taken together indicate obvious differences be sucrose and 0.1 mw CaCI2 adjusted to pH 8.0 with 1 M Tris buffer and tween the low- and high-metastatic cells, Eb and ESb, in the homogenized with a glass homogenizer. The homogenate was spun at expression of receptor sites for lectins with specificity for GalNAc 1500 rpm for 10 min. The supernatant was collected and stored at -70Â°. residues (VV, SBA, and HP), which are expressed on Eb but not Asialofetuin and asialoagalactofetuin were prepared from fetuin (Grade ESb cells. On the other hand, receptor sites for PNA, which has III; Sigma Chemical Co., Munich, Federal Republic of Germany) according specificity for D-Gal-1-3-D-GalNAc configurations, were better to standard procedures (32, 33). Ovalbumin (crystallized 3 times) was expressed on ESb than on Eb cells. Removal of SA by neura purchased from Miles Laboratories, and RNase A was from Sigma. Assays for sialyltransferase (5), galactosyltransferase (6), fucosyltrans- minidase treatment led to the uncovering of binding sites for ferase (7), and N-acetylglucosaminyltransferase (5) were performed es SBA and VV by ESb cells and of terminal o-Gal-1-3-o-GalNAc sentially as described in the literature. [14C]CMP SA, [14C]UDP galactose, groups on Eb cells detectable by the respective specific lectins. [I4C]GDP-fucose, and [14C]UDP-W-acetyl-D-glucosamine were obtained We have recently isolated a plastic-adherent variant line from from NEN (Dreieich, Federal Republic of Germany). the high-metastatic ESb cell (ESb-M) which was found to have lost its metastatic potential while still being tumorigenic in normal syngeneic hosts (14). The variant retained most of its ESb- RESULTS derived antigenic and biochemical characteristics, but showed Staining of Eb and ESb Tumor Cells with FITC-conjugated increased expression of binding sites for the lectins VV and SBA, Lectins. FITC-conjugated lectins with different sugar specificities very similar to the characteristics of the low-metastatic parental (Table 1) were used to study the expression of corresponding line Eb. While such sites were masked by SA on metastatic ESb binding sites on the standard Eb and ESb tumor cells. Although cells, they became unmasked on ESb-M cells. Furthermore, cloned sublines of both tumor cell populations were available, metastatic revertants (ESb-MR) from ESb-M cells isolated from we intentionally studied here the noncloned populations to get the spleen (Esb-MSP) or the brain (ESb-MBR) of tumor-bearing an impression of the overall differences. As negative controls, mice were again found to have the respective lectin binding sites we used lectin binding values in the presence of 0.2 M inhibitory masked by SA. There was, thus, within the lineage Eb â€”Â»ESb sugars. The results obtained for ESb and Eb cells with or without â€”Â»ESb-Mâ€”Â»ESb-MR,a correlation between the phenomenon neuraminidase treatment are shown in Chart 1. The lectins VV, described here, i.e., exposure or masking of SBA receptor sites SBA, and HP stained the metastatic ESb tumor cells only weakly and overall metastatic capacity (Table 2). (close to background controls), while the parental line Eb was Staining of MDAY-D2 Tumor Cells and Its Wheat Germ able to bind these lectins (Chart 1). The failure of these lectins Agglutinin Selected Mutants with VV and SBA. In order to to bind to ESb cells was apparently due to a masking of receptor evaluate whether the relation between exposure or blocking of sites by SA, since treatment of the cells with neuraminidase SBA receptor sites and metastatic capacity was a peculiarity of uncovered these sites and led to staining of ESb cells to a similar the Eb/ESb tumor system or was perhaps of more general degree as that for Eb cells (Chart 1). There was, however, significance, we investigated another tumor system consisting virtually no staining observed with HP, even when neuraminidase of high- and low-metastatic related tumor lines. The parental treatment was performed. This is in agreement with the previous high-metastatic tumor line MDAY-D2, the nonmetastatic WGAR observation that T130, the major HP binding glycoprotein (2), is mutant, and 2 metastatic revenant lines were compared for their

Table 1 Sugar specificities of lectins used in this study HP D-GalNAc-a-1 -Â» W D-GalNAc-Â«-1â€”3-D-Gal-/3-1-Â» Glycine max (SBA) D-GalNAc-a-1 -Â»3-o-Gal-/3-1 -> 3-D-GlcNAc Ulex europeus L-Fuc-a-1 â€”2-D-Gal-tf-1 â€”4-D-GlcNAc-/3-1 - Ricinus communis D-Gal-/3-l-Â» Arachis hypogea (PNA) D-Gal-/3-1 -> 3-D-GalNAc Conavalia enziformis (concanavalin A) D-Man-a-1 -Â»

NOVEMBER 1983 5139

W SBA HP PNA

XXÂ» XXÂ» 1000 1000

7.7 0.3 5.5 l 2.2

â€¢4-*8 J o - Ã¼

15.1 16 6.5 31.7 SÃ¬ - co LU -

80.5 62'Â° 13.8 â€¢oto8 '

Å“ &LUZ

2.8 0.1 8.9 0.1

61.5 61.1 33.7 1.1

a 83.7 71.0 56.8 70.5 'c -

200 600 1000 200 600 1000 200 600 1000 200 600 1000 X 1 GREEN FLUORESCENCE Chart 1. Histograms from cytofluorographic analysis of tumor cells stained with the indicated FITC-conjugated lectins (Medac, Hamburg, Federal Republic of Germany). Background controls were done by incubating cells and lectins in the presence of 0.2 M inhibitory sugars (GalNAc for VV. SBA, and HP; D-Gal for PNA and Ricinis communis agglutinin; methylmannoside for concanavalin A; and L-Fuc for Ulex europaeus agglutinin). Three x 104 cells were analyzed for each histogram, and the

5140 CANCER RESEARCH VOL. 43

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1983 American Association for Cancer Research. Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1983 American Association for Cancer Research. P. Altevogt et al.

Table 2 Binding of FITC-conjugated lectins W and SBA to ESb-derived tumor cell lines Cells were stained and analyzed by flow cytofluorography as described in the legend to Chart 1. bindingVV24.6 of cells A*93.189.4 CellsESbESb-M capacityHigh

Low 74.9 63.4 ESb-MSP High 28.7 33.9 93.7 ESb-MBRControls High% 38.714.0 44.29.0 87.21

in presence of sugarMetastatic (0.2 M GalNAc)SBA30.2 (0.2 M GalNAc)Concanavalin1.0 (0.2 M n-methylmannoside) " Binding of concanavalin A to internal sugar moieties was more uniform and could be inhibited in the presence of Â«-methylmannoside.

Table 3 Binding of FITC-conjugated lectins VV and SBA to MDAY-D2-derived tumor cell lines Cells were stained and analyzed by flow cytofluorography as described in the legend to Chart 1. of cells binding W of cells binding neuraminidase SBAneuraminidasetreatmentÂ»1.8 the treatment2.886.5 presence of the preseneÂ« CellsMDAY-D2 capacityHigh sugar1.8 of sugar0.5

MDW40 Low 17.7 94.7 5.6 80.2 86.6 MDW40M1 High 6.4 11.3 75.0 1.5 5.8 82.5 MDW40M4Metastatic HighIn 16.2% 27.6+94.3 90.5In 2.1% 3.1+81.1 84.7

Table 4 enzyme activities using standard procedures (5-8) (Table 5). Neuraminidase-releasable SA of murine tumor cell lines used in this study Table 5 shows that A/-acetylglucosaminyltransferase and sialyl- CellsMDAY-D2MDW40MDW40M1MDW40M4EbESbESb-MMetastaticcapacityHighLowHighHighLowHighLownmolprotein1.91SA/mg cell transferase activities were, respectively, 5- or 2-fold higher in Eb (5)"0.59Â±0.35a than in ESb tumor cells. Activities of fucosyltransferase A and B (4)1 Â±0.13 and galactosyltransferases were only slightly different. (2)1.69.89 Â±0.06 (3)1.08Â±0.12 These data clearly document changes in glycosyltransferase activities between Eb and ESb cells. (2)1.42+ 0.19 (3)1.45Â±0.10 Â±0.01 (2) DISCUSSION " Mean Â±S.D. 6 Numbers in parentheses, number of experiments performed. Changes in the metastatic behavior of defined tumor cell lines have been observed to be associated with changes in cell surface binding properties to the lectins VV and SBA. The results, glycosylation (9,13). An important role of surface carbohydrates presented as the percentage of positive cells, are shown in Table was also supported by the effect of tunicamycin on metastatic 3. The parental metastatic line MDAY-D2 and the wheat germ arrest properties (16). In this particular context, SA has received agglutinin-sensitive revenants, MDW40M1 and MDW40M4, increased attention recently (36). could not be stained significantly above the background controls In the present study, we have analyzed cell surface carbohy with either of the 2 lectins. They were, however, readily stained drates by means of lectin binding in 2 different tumor systems if the cells were pretreated with neuraminidase. In contrast, the consisting of various tumor sublines differing greatly in metastatic nonmetastatic WGAR MDW40 cells stained brightly without neur potential. The most consistent differences in lectin binding prop aminidase treatment. erties between the pairs of high- and low-metastatic tumor lines Determination of SA Content of the Tumor Cell Lines. The investigated (i.e., Eb versus ESb; ESb versus ESb-M; ESb-M results presented indicated similar differences in VV and SBA versus ESb-MR; MDAY-D2 versus MDW40; and MDW40 versus lectin staining between high- and low-metastatic tumor lines of MDW40M1) were found with the D-GalNAc-specific lectins, VV 2 different tumor systems. Since the differences found could be and SBA. Both lectins were found to bind directly only to the affected by neuraminidase pretreatment, we determined the total low-metastatic tumor lines, while the high-metastatic lines re amount of neuraminidase-releasable SA for all of the lines stud quired neuraminidase pretreatment in order to expose the re ied. As shown in Table 4, cells from the MDAY-D2 lineage spective binding sites. showed a 3-fold reduction of SA for the nonmetastatic MDW40 In a recent study, Kaladas ef al. (17) have investigated the fine line, while the revenants showed values similar to those of the specificity of the interaction of VV lectin to carbohydrates. parental line. In the case of the Eb, ESb, and ESb-M cells, the The purified lectin agglutinated type A human erythrocytes pref values for neuraminidase-accessible SA differed only slightly. At erentially. Inhibition assays with various monosaccharides, gly- least in the Eb/ESb system, there was no correlation between cosides, and oligosaccharides indicated that the VV lectin is metastatic potential and net amount of neuraminidase-releasable specific for terminal, nonreducing Â«-linked D-GalNAc (17) resi SA. dues. Among the disaccharides tested, o-GalNAc-Â«-1-3-D-Gal Determination of Glycosyltransferase Activities of Eb and was most active in inhibition, and the importance of the Â«-1-3- ESb Tumor Cells. Altered levels of glycosyltransferase activities linkage was indicated (17). In contrast to VV, SBA is known to of ESb tumor cells could possibly account for the noted differ react strongly with o-galactose as well as with o-GalNAc (25) ences in glycosylation. Therefore, we determined some of these and has no anomeric specificity (17). The question of which

5142 CANCER RESEARCH VOL. 43

Table 5 Glycosyltransferase activities of the Eb and ESb tumor cells Enzyme activities (cpm/mg protein)

activitiesSialyltransferaseGalactosyltransferaseFucosyltransteraseAnalyzed enzyme acid)CMP(["Cjsialic ,400 Â±618a433 Â±87466 acid)UDP(["CJsialic Â±157,753 Â±2810,992 (ÃŒ14C]galactose)UDP Â±85650 Â±178775 (['4C)galactose)GDP Â±1825,220 Â±2020,260 AFucosyltransferase ([14C]fucose)GDP Â±2111000 Â±1201 (r14C]fucose)GDP Â±2117,950 ,400 Â±2513,31 B/v-Acetylglucosaminyltransfer-ase-1N-Acetylglucosaminyltransfer-ase-2SubstrateCMP(['4C)fucose)GDP Â±30450 6Â±82550 (|'4C]fucose)UDP Â±1020,925 Â±124,1 ([14C)NAcGlc)UDP Â±159320 75Â±39.1,125

(['4C]NAcGlc)UDP Â±405,000 ([14C]NAcGlc)UDP Aâ€”Eb61 Â±18375 Â±32550

(["CJNAcGIc)AcceptorAsialofetuinOAsialoagalactofetuinâ€”Asialofetuinâ€”Asialoagalactofetuinâ€”Ovalbuminâ€”RNaseÂ±38ESb32,957 Â±13ESb/Eb0.541.410.800.740.200.23 a Mean Â±S.D. 6 â€”,assays performed in absence of exogenous acceptor; the values obtained include background plus possible contribution of endogenous acceptor. tumor cell surface structures were involved in the binding of VV neuraminidase-accessible cell surface SA was quite similar. and SBA has to be discussed in light of these specificity data. Therefore, the difference in sialylation of the receptor sites for Terminal GalNAc residues, which could act as potential binding VV and SBA in this tumor system was most likely due to sites for SBA and VV, have not been observed to be sialylated differences in the exact positioning of SA on the various glyco- in any glycolipid or W-glycosidically linked protein carbohydrate conjugates of the cell surface. In the MDAY-D2 tumor system, side chain. 0-Glycosidically linked mucin-type carbohydrate side the low-metastatic lectin-resistant variant MDW40 had only chains have been observed to be sialylated on galactose residues about one-third of the amount of neuraminidase-accessible SA or on the innermost GalNAc moiety. Biochemical studies are of the corresponding metastatic lines (i.e., the parental line and presently being performed on the tumor cell surface carbohy 2 metastatic revertants). In this system, therefore, increase in drates with serve as receptor sites for VV and SBA, on their metastatic capacity was associated both with an increase in the linkage type, and on the nature of the respective glycoconju- total amount of cell surface sialylation and in the site-specific gate(s). sialylation of SBA and VV receptor sites. From these and other Interestingly, it was shown by Kimura and Wigzell (20) that findings (36), it seems that the receptor sites for SBA or VV on cytotoxic mouse T-lymphocytes display a cell surface glycopro- tumor cells can have a negative impact on their metastatic tein, T145, to which VV has an exquisite specificity. A similar capacity. glycoprotein could be detected after neuraminidase treatment How could the differences between Eb and ESb cells in and selective labeling by the galactose oxidase/NaBH4 technique expression of specific lectin binding sites be explained? Relevant on some cloned ESb tumor cells (1). The T145 glycoprotein to this context may be our recent observations on hepatocyte- among putative others might be a candidate for the interaction tumor cell interactions in vitro, where we found that only ESb of VV with neuraminidase-treated ESb tumor cells. tumor cells and not Eb cells were able to form spontaneous In the 2 tumor systems studied, the binding of VV and SBA rosettes with the hepatocytes (29). Eb cells required treatment correlated with the metastatic capacity of the tumor sublines. In with neuraminidase in order to exhibit similar rosette formation the B16 melanoma system, however, which, in contrast to (29). A detailed molecular analysis of this interaction allowed the MDAY-D2 and Eb/ESb cells, is plastic-adherent, binding of iodi- identification of a lectin-like protein on hepatocytes (HBP), which nated SBA did not correlate with the metastatic capacity of the seemed to recognize terminal Gal-GalNac residues and to be cells (27). Therefore, it will be interesting to extend lectin binding similar to PNA in specificity." As can be seen from Chart M3, this studies to other tumor cells with differences in metastatic poten lectin, in contrast to VV or SBA, reacted more strongly with ESb tial and adhesion properties in order to evaluate a possible than with Eb cells. Thus, it appears as if the transition from the general significance of our findings. low-metastatic Eb cells to their high-metastatic variant was Little is known at present about the possible physiological role associated with a change in the sialylation of specific binding of the carbohydrate side chains which serve as specific lectin sites. While Eb cells expressed binding sites for VV and SBA receptor sites for VV or SBA. It is conceivable that such sites and had the binding sites for HBP and PNA masked by sialylation, might mediate cellular adhesion to certain substrates, such as ESb cells had the binding sites for VV and SBA sialylated and fibronectin, laminin, or collagen. In support of this assumption, instead expressed binding sites for HBP and PNA. A difference we have recently found that the low-metastatic variant MDW40, in the positioning of terminal SA on high- and low-metastatic which showed a much better expression of receptor sites for VV tumor cells could thus have a dual effect: a blocking of certain and SBA than did the metastatic parental line MDAY-D2 (Table binding sites and an unblocking of others. This could have a 3), also showed a much better attachment to basement mem pronounced effect on the biological behavior of the respective brane collagen type IV and fibronectin than to laminin. Further cancer cells (10,14, 29) and could lead to altered charge distri more, neuraminidase treatment of the metastatic MDAY-D2 tu bution patterns of metastatic versus nonmetastatic cells as was mor cells selectively enhanced their attachment to collagen IV 4 R. Cheingsong-Popov, P. Robinson, P. Altevogt, and V. Schirrmacher. A mouse and fibronectin (10). hepatocyte carbohydrate-specific receptor and its interaction with metastasizing In the Eb/ESb/ESb-M tumor system, the total amount of tumor cells. Int. J. Cancer, in press, 1983.

NOVEMBER 1983 5143

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1983 American Association for Cancer Research. P. Altevogt et al. shown Â¡nthe B16 melanoma system (26). 162, 1978. A further important question is the possible role of glycosi- 8. Chatterjee, S. K., Kim, U., and Bielat, K. Plasma membrane-associated en zymes of mammary tumors as the biochemical indicators of metastasizing dases and glycosyltransferases in the modulation of expression capacity. Analysis of enriched plasma membrane preparations. Br. J. Cancer, of carbohydrate binding sites. A proper interpretation of the 33: 15-22. 1976. 9. Dennis, J., Donaghue, T. P., Florian, H., and Kerbel, R. S. Apparent reversion results on the glycosyltransferase activities of Eb and ESb cells of stable in vitro genetic markers detected in tumor cells from spontaneous (Table 4) depends on the carbohydrate side chain structure of mÃ©tastases.Nature (Lond.), 292. 242-245, 1981. the acceptor proteins used in these assays. This seems of 10. Dennis, J., Waller, C., Timpl, R., and Schirrmacher, V. Surface sialic acid reduces attachment of metastatic tumour cells to collagen and fibronectin. particular importance, because it appears that there exist gly Nature (Lond.), 300: 274-276, 1982. cosyltransferases for the same sugar with different substrate 11. Dennis, J. W., Donaghue, T. P., and Kerbel, R. S. Membrane-associated alterations detected in poorly tumorigenic lectin-resistant variant sublines of a fine specificity (3, 4, 24). Fetuin, asialofetuin, and asialogalacto- highly malignant and metastatic murine tumor. J. Nati. Cancer Inst., 66: 129- fetuin were used as acceptor proteins for the determination of 139, 1981. sialyl-, galactosyl-, and fucosyltransferase activities. It is known 12. Finne, J., Burger, M. M., and Prieels, J-P. Enzymatic basis for lectin-resistant phenotype: increase in a fucosyltransferase in mouse melanoma cells. J. Cell from earlier studies by Spiro (32, 33) on the structure of the Biol., 92: 277-282,1982. carbohydrate units of fetuin that the O-glycosidically linked ones 13. Finne, J., Tao, T. W., and Burger, M. M. Carbohydrate changes in glycoproteins have a terminal sequence of SA-o-Gal-GalNAc-(SA)-Ser(Thr), of a poorly metastasizing wheat germ agglutinin-resistant melanoma clone. Cancer Res., 40. 2580-2587, 1980. while the A/-glycosidically linked ones have a terminal sequence 14. Fogel, M., Altevogt, P., and Schirrmacher, V. Metastatic potential severely of SA-D-Gal-GlcNAc4-Man3-GlcNAc-GlcNAc-Asn. Accordingly, altered by changes in tumor cell adhesiveness and cell surface sialylation. J. Exp. Med., 757: 371-376, 1983. asialofetuin expresses numerous terminal galactosyl-residues 15. Gottlieb, C., Baenzinger, J., and Kornfeld, S. Deficient uridine diphosphate-N- and thus can be assumed not to test for all sialyltransferases acetylglucosamine: glycoprotein W-acetylglucosaminyltransferase activity in a but perhaps predominantly for galactosyl-specific ones. The clone of Chinese hamster ovary cells with altered surface glycoproteins. J. Biol. Chem., 250: 3303-3309,1974. finding of a higher sialyltransferase activity in Eb cells (possibly 16. Irimura, T., Gonzales, R., and Nicolson, G. L. Effect of tunicamycin on B16 specific for galactosyl residues) could thus account for the higher metastatic melanoma cell surface glycoproteins and blood-borne arrest and degree of sialylation of terminal ÃŸ-galactose structures as rec survival properties. Cancer Res., 47: 3411-3418,1981. 17. Kaladas, P. M., Kabat, E. A., Kimura, A. K., and Ersson, B. The specificity of ognized by HBP or PNA. This is also in agreement with the the combining site of the lectin from Vicia villosa seeds which reacts with observation that Ricinus commuais agglutinin receptor sites on cytotoxic T lymphoblasts. Mol. Immunol.. 73: 969-977,1981. 18. Kerbel, R. S., Florian, H., Man, M. S., Dennis, J., and McKencie, l. F. C. Eb cells are more sensitive to neuraminidase treatment than are Carcmogenicity of tumor cell populations: origin of a putative H-2 isoantigenic those on ESb cells (Chart 10). Similarly, the higher activity of loss variant tumor. J. Nati. Cancer Inst., 64:1221-1230,1980. galactosyltransferase in ESb-type cells could contribute to a 19. Kerbel, R. S., Twiddy, R. R., and Robertson, D. M. Induction of a tumor with greatly increased metastatic growth potential by injection of cells from a low better expression of HBP or PNA binding sites on these cells. metastatic H-2 heterozygous tumor cell line into an H-2 incompatible parental The strongly reduced activities of NAc-glucosaminyltransferases strain. Int. J. Cancer, 22: 583-594, 1978. in ESb cells could perhaps indicate a less complete and complex 20. Kimura, A. K., and Wigzell, H. Cell surface glycoproteins of murine cytotoxic T lymphocytes. I. T145, a new cell surface glycoprotein selectively expressed biosynthesis of carbohydrate side chains as compared to Eb- on Ly 1-2* cytotoxic T lymphocytes. J. Exp. Med., 747: 1418-1434,1978. type cells, but further experimentation is required before such a 21. Mannino, R. J., and Burger, M. M. Growth inhibition of animal cells by succinated concanavalin A. Nature (Lond.), 256: 19-22, 1975. conclusion might be drawn. 22. Meager, A., Ungkitchanukit, A., Nairn, R., and Hughes, R. C. Ricin resistance Based on the results presented here and from previous work in baby hamster kidney cells. Nature (Lond.), 257:137-139,1975. 23. Narasimhan, S., Stanley, P., and Schachter, H. Control of glycoprotein biosyn on glycosylation differences between Eb and ESb cells (1, 31 ), thesis: lectin-resistant mutant containing only one of two distinct A/-acetylglu- we propose: (a) that cell surface carbohydrate changes take cosaminyltransferase activities present in wild type Chinese hamster ovary place during tumor progression; (o) that these changes are not cells. J. Biol. Chem., 252. 3926-3933, 1977. 24. Paulson, J. C., Rearick, J. I., and Hill, R. L. Purification of a sialyltransferase random but show certain patterns detectable with lectins; and from bovine colostrum by affinity chromatography on CDP-agarose. J. Biol. (c) that glycosyltransferases and, in particular, sialyltransferases Chem., 252: 2363-2371, 1977. play an important role in these changes, leading to the masking 25. Pereira, M. E. A., Kabat, F. A., and Sharon, N. Immunochemical studies on the specificity of soybean agglutinin. Carbohydr. Res., 37: 89-102, 1974. of some binding sites and the exposure of others. 26. Raz, A., Bucana, C., McLellan, W., and Fidler, I. J. Distribution of membrane anionic sites on B16 melanoma variants with differing lung colonizing potential. REFERENCES Nature (Lond.), 284: 363-364,1980. 27. Raz, A., McLellan, W., Hart, I. R., Bucana, C. D., Hoyer, L. C., Sela, B. A., 1. Altevogt, P., Kumick, J. T., Kimura, A. K., Bosstet, Â«.,and Schirrmacher, V. Dragsten, P., and Fidler, I. J. Cell surface properties of B16 melanoma variants Different expression of Lyt differentiation antigens and cell surface glycopro- with differing metastatic potential. Cancer Res., 40. 1645-1651, 1980. teins by a murine T lymphoma line and its highly metastatic variant. EUT. J. 28. Reading, C. L., Betoni, P. N., and Nicolson, G. L. Selection and in vivo Immunol., 12: 300-307, 1982. properties of lectin-attachment variants on malignant lymphosarcoma cell lines. 2. Axelsson, B., Kimura, A., Hammarstrom, S., Wigzell, H., Nilsson, K., and J. Nati. Cancer Inst., 64: 1241-1249, 1980. Mellstedt, H. Helix pomatia A hemagglutinin: selectivity of binding to lympho 29. Schirrmacher, V., Cheingsong-Popov, R., and Amheiter, H. Hepatocyte-tumor cyte surface glycoproteins on T cells and certain B cells. Eur. J. Immunol., 8: cell interaction in vitro. I. Conditions for rosette formation and inhibition by anti 757-764,1978. H-2 antibody. J. Exp. Med., 757. 984-989,1980. 3. Bartholomew. B. A., Jourdian, G. W., and Roseman. S. The sialic acids: XV. 30. Schirrmacher, V., Fogel, M., Russmann, E., Bosslet, K., Altevogt, P., and Transfer of sialic acid to glycoproteins by a sialyltransferase from colostrum. Beck, L. Antigenic variation in cancer metastasis: immune escape versus J. Biol. Chem., 248: 5751-5762,1973. immune control. Cancer Met. Rev., 7: 241-274, 1982. 4. Carlson, D. M., Jourdian, G. W., and Roseman, S. The sialic acids: XIV. 31. Schwartz, R. Ph.D. Dissertation, University of Braunschweig, 1982. Synthesis of sialyl-lactose by a sialyltransferase from rat mammary gland. J. 32. Spiro, R. G. Periodate oxidation of glycoprotein fetuin. J. Biol. Chem., 239: Biol. Chem., 248: 5742-5750,1973. 567-573, 1964. 5. Chatterjee, S. K. Glycosyltransferases in metastasizing and non-metastasizing 33. Spiro, R. G. Analysis of sugars found in glycoproteins. Methods Enzymol., 8: rat mammary tumors and the release of these enzymes in the host sera. Eur. 3-26, 1966. J. Cancer, 75: 1351-1356, 1979. 34. Tao, T. W., and Burger, M. M. Non-metastasizing variants selected from 6. Chatterjee, S. K., and Kim, U. Galactosyltransferase activity in metastasizing metastasizing melanoma cells. Nature (Lond.), 270: 437-438,1977. and non-metastasizing rat mammary carcinoma and its possible relationship 35. Warren, L. The thiobarbituric acid assay of sialic acids. J. Biol. Chem., 243: with tumor cell surface antigen shedding. J. Nati. Cancer Inst., 58: 273-280, 1971-1979,1959. 1977. 36. Yogeeswaran. G., and Salk, P. L. Metastatic potential is positively correlated 7. Chatterjee, S. K., and Kim, U. Fucosyltransferase activity in metastasizing and with cell surface sialylation of cultured murine tumor cell lines. Science (Wash. non-metastasizing rat mammary carcinomas. J. Nati. Cancer Inst., 67: 151- D.C.), 272:1514-1516,1981.

5144 CANCER RESEARCH VOL. 43

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1983 American Association for Cancer Research. Different Patterns of Lectin Binding and Cell Surface Sialylation Detected on Related High- and Low-Metastatic Tumor Lines

Peter Altevogt, Mina Fogel, Rachanee Cheingsong-Popov, et al.

Cancer Res 1983;43:5138-5144.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/43/11/5138

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/43/11/5138. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.