[CANCER RESEARCH 40, 1645-1651, May 1980]

Cell Surface Properties of B16 Melanoma Variants with DifferingMetastatic Potential1

A. Raz,2 W. L. McLellan, I. R. Hart, C. D. Bucana, L. C. Hoyer, B-A. Sela, P. Dragsten, and I. J. Fidler

Cancer Biology Program. National Cancer Institute Frederick Cancer Research Center, Frederick, Maryland 21701 [A. R., W. L. M.. I. R. H., C. D B., L. C. H., l. J.F.]; Department of Biophysics, The Weizmann Institute of Science, Rehovot. Israel ¡B-A. SJ; and National Cancer Institute, Laboratory of Theoretical BiologyBethesda, Maryland 20205 [P. D.]

ABSTRACT

Several cell surface properties of three B16 melanoma variant lines, which exhibit differing lung-colonizing capacities,have been examined. Equal numbers of cells from the B16-F10Lr and B16-F1 lines, injected i.v. into syngeneic mice,

produced significantly fewer lung nodules than did cells of theB16-F10 line. These wide differences in biological behavior

could not, however, be ascribed to major qualitative differencesin the cell surface properties studied.

The 3 cell lines displayed similar quantitative and qualitativepatterns of exposed surface proteins. The quantity of radioactive labeling of the major sialoglycoprotein (78,000 dallons)was inversely proportional to the capacity of cells to producelung tumor colonies. Lectin-binding assays revealed that B16-F10Lr cells bound 50% less concanavalin A than did the B16-

F1 or B16-F10 cells; the B16-F10 cells bound 40% less wheatgerm agglutinin than did the B16-F1 or B16-F10Lr cells. Levelsof 5'-nucleotidase activity were lower in B16-F10 cells than inB16-F1 and B16-F10Lr cells. At the same time, there was no

reduction in levels of the cytoplasm-associated enzyme, acid

phosphatase. No differences in membrane lipid fluidity amongthe cell lines were detected by either fluorescent polarizationor photobleaching techniques. The kinetics of adhesion of B16-F10Lr and B16-F10 to a variety of substrates were very similar.

The present data suggest that wide differences in metastaticbehavior of B16 melanoma cells need not be correlated withmajor qualitative changes in a single cell surface biochemicalor morphological characteristic.

INTRODUCTION

The process of metastasis involves the release of cells fromthe primary tumor, dissemination to distant sites, arrest in themicrocirculation of organs, extravasation and infiltration intothe stroma of those organs, and survival and growth into newtumor colonies. The outcome of metastasis depends on bothhost factors and tumor cell properties, which may vary amongtumor systems (15). Since the interaction of metastatic cellswith their environment is mediated to a considerable extent bythe cell surface, it seems likely that cell surface propertiescould influence the metastatic potential of tumor cells. Changesin the cell surface that accompany neoplastic transformationhave been studied in great detail. In fact, almost all characteristics of the transformed cell membrane appear altered when

1 Research sponsored by the National Cancer Institute under Contract NO1-

CO-75380 with Litton Bionetics, Inc.2 To whom requests for reprints should be addressed.

Received August 20, 1979; accepted February 6, 1980.

compared to those of the untransformed counterpart (21, 24,33, 47). The relationship of any specific cell surface change tothe outcome of experimental metastasis, however, is unknown.

In order to identify those properties of tumor cells that allowthem to interact favorably with their host and to establishmetastatic growths, it is advantageous to study variants of thesame tumor with differing capacities to form experimental métastases. The isolation of tumor variants can be achieved byseveral methods. The first involves isolation and propagationof variant lines from métastases.In such diverse murine tumorsystems as melanoma (12), fibrosarcoma (44), methylcholan-threne-induced sarcoma (26), lymphosarcoma (7), Lewis lung

carcinoma (18), and lymphoma (27), variants with differentmetastatic potentials have been selected. A second method forisolation of tumor variants involves in vitro selection of cells forspecific properties, such as detachment from a monolayer (5),resistance to lysis by syngeneic lymphocytes (14), and resistance to lectin-mediated toxicity (46), and then to determine

whether the selection affects metastatic behavior in vivo. Athird more recent and complementary approach has been toestablish in vitro a large number of clones or lines from parenttumors such as the B16 melanoma (6, 12, 13, 16), a UV-

induced fibrosarcoma (28), a mammary tumor (10), and amethylcholanthrene-induced fibrosarcoma (43). Using this ap

proach, the behavior in vivo and in vitro of such clones can bestudied.

In recent years, we have concentrated on the B16 melanoma,syngeneic to the C57BL/6 mouse, as a tumor system. Bymanipulating various host and tumor variables, we have attempted to determine the relative contributions of several factors to the survival and growth of blood-borne tumor cells. Wehave described the selection in vivo (12) and in vitro (14) ofvariant sublines of the B16 melanoma with different capacitiesfor experimental metastasis. Several biochemical differencesbetween the cells of low and high lung-colonizing potential

have been reported. Among these were electrophoretic mobility and surface glycoproteins (19). However, comparative analysis of surface proteins labeled by lactoperoxidase iodinationand glycopeptides released by mild trypsin-pronase digestion

failed to demonstrate any significant differences between highland low-metastatic B16 cell variants (34, 48).

The present report concerns our analysis of the functional,structural, morphological, and dynamic properties of the cellmembrane components and their correlation with the lung-colonizing behavior of 3 variants of the B16-F series. The

results indicate that no qualitative differences in major components of the cell membrane among the lines can be demonstrated. However, subtle quantitative and perhaps cumulativechanges in the cell surface may contribute to the success orfailure of blood-borne tumor cell survival and growth in vivo.

MAY 1980 1645

on May 21, 2020. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

A. Raz et al.

MATERIALS AND METHODS

Animals

Specific-pathogen-free C57BL/6 mice were obtained fromthe Animal Production Area of the Frederick Cancer ResearchCenter. For each experiment, all mice were age and sexmatched.

Tumor Cell Variants

The B16-F1 cell line (low incidence of lung colonization) wasderived from pulmonary métastasesproduced by i.v. injectionof B16 melanoma cells obtained from The Jackson Laboratory,Bar Harbor, Maine. The B16-F10 cell line (high incidence of

lung colonization) was selected by successive passaging oflung tumor colonies for 10 in vivo-in vitro selections as described in detail previously (12). B16-F10Lr (lymphocyte resist

ant, low incidence of metastasis to the lung) is a variant cellline selected from B16-F10, after tumor cell monolayers were

subjected to cytotoxic lymphocytes 6 times in vitro (14).

Tumor Cultures

All cultures were maintained on plastic in Eagle's minimal

essential medium supplemented with 10% fetal bovine serum,sodium pyruvate, nonessential amino acids, L-glutamine, and2-fold vitamin solution (Flow Laboratories, Inc., Rockville, Md.).Antibiotics were not included in the medium for routine maintenance of the cultures. Cell cultures were grown at 37°in a

humidified atmosphere containing 5% CCv The cell lines wereexamined for and found to be free of Mycoplasme and thefollowing murine viruses: reovirus type 3; pneumonia virus ofmice; K virus; Theiler's virus; Sendai virus; minute virus of

mice; mouse adenovirus; mouse hepatitis virus; lymphocyticchoriomeningitis virus; ectromelia virus; and lactate dehydro-

genase virus (Microbiological Associates, Bethesda, Md.). Thecell lines were grown from frozen stocks and were maintainedin continuous cultures for periods not exceeding 60 days. Thelung colonization potential of the 3 variant lines was evaluatedseveral times during the studies, as described below. The celllines were coded for all in vitro assays.

Experimental Pulmonary Metastasis

Cells from semiconfluent monolayers were harvested with ashort (2-min) treatment of 2 rriM EDTA in a Ca2+ and Mg2+-freePBS,3 pH 7.2. The cells were washed in PBS and pipetted

gently to dissociate all cell clumps. Only cell suspensions ofmore than 90% viability, as measured by trypan blue exclusion,and those free of all aggregates were used for in vivo or in vitrostudies. Unanesthetized mice were given i.v. injections of 0.2ml containing 50,000 viable single tumor cells suspended inHanks' balanced salt solution via the tail vein. All mice were

killed 18 days after tumor cell injection, and their lungs wereremoved, rinsed in water, and fixed in formalin. The number oflung tumor colonies was determined by counting surface métastases under a dissecting microscope, since the majority of

3 The abbreviations used are: PBS. phosphate-buttered saline (0.01 M phosphate buffer, pH 7.2-0.15 M NaCI); SBA, soybean agglutinin; UEA1, Ulex euro-peus agglutinin I; Con A. concanavalin A; WGA. wheat germ agglutinin; WBA,wax bean agglutinin; DPH, 1,6-diphenyl-1,3,5-hexatriene; DM,3,3'-dioctadecylin-

docarbocyanine.

such experimental métastases in mice are located near thelung surface (11 ). Métastaseswere counted in a blind fashionby 2 observers. All extrapulmonary métastaseswere also recorded.

Lectins

The following lectins were radioactively labeled by the 125I-

chloramine-T method (23): SBA, UEA1 (Vector Laboratory,Burlingame, Calif.), Con A, WGA (Miles Yeda, Rehovot, Israel),and WBA. The radioactively labeled lectins were separatedfrom free iodine by gel filtration on a Bio-Gel P-2 column.Specific radioactivity of the lectins ranged from 3 to 10 x 106

cpm/mg protein.For quantitative studies with radioactively labeled lectins, 0.5

x 106 tumor cells were exposed to increasing concentrationsof a given lectin for 30 min at 24°. The specific binding of a

lectin to the cell surface was calculated by subtracting theamount bound in the presence of a specific inhibitor (to lectinbinding) from the amount of lectin bound in the absence of itsinhibitor. The specific inhibitors were 0.1 M a-methyl-D-man-nopyranoside for Con A, 0.1 M N-acetylglucosamine for WGA,0.1 M A/-acetylgalactosamine for SBA, 0.1 M L-fucoseforUEA1,

and 0.1% fetuin for WBA. Lectin binding in the presence ofspecific inhibitors constituted less than 10% of total binding.

Neuraminidase Treatment

B16 cells suspended in PBS were incubated for 30 min at37°with protease-free Vibrio cholera neuraminidase (50 units/

ml) purchased from Calbiochem, La Jolla, Calif., and thenwashed in buffer to remove the free sialic acid and the enzyme.When comparison between untreated and neuraminidase-

treated cells was desired, the untreated cells were incubatedin PBS for the same duration as the enzyme treatment. No lossof cell viability could be detected following neuraminidasetreatment.

Surface Labeling of Glycoproteins

Galactose residues of cell surface glycoproteins were radio-

labeled before or after neuraminidase treatment. Cells wereincubated with 50 jug of galactose oxidase for 180 min at 37°

and then centrifuged, washed in, and reduced with 0.5 mCiNaB3H4 (New England Nuclear, Boston, Mass.) for 30 min at24°. Cells were washed and processed for electrophoresis.

For surface labeling of sialic acid residues of glycoproteins, 2x 106 cells suspended in 1 ml of phosphate-buffered NaCI

(0.13 M NaCI in 0.02 M potassium phosphate, pH 7) wereincubated with 5 row NalO4 at 4°for 15 min (19); 0.2 ml of 0.1

M glycerol was then added, and the cells were incubated foran additional 10 min. After being washed with the buffer, thesuspensions were reduced with [3H]borohydride, and the cells

were processed for electrophoresis. Labeling procedures didnot lead to decreased cell viability.

Electrophoresis

The labeled cells were solubilized with 1% Triton X-100, 0.4M KI-0.05 M Tris-HCI, pH 9.0, and centrifuged at 10,000 x g

for 15 min to remove insoluble material. An aliquot was resolvedby electrophoresis separation on a 5 to 20% gradient of sodiumdodecyl sulfate-polyacrylamidegel (29). Fluorography wascar-

1646 CANCER RESEARCH VOL. 40

on May 21, 2020. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Surface Properties of Metastatic Cells

ried out by the method of Bonner and Laskey (3). Gels werefixed, stained and destained, and then incubated with 22%dimethyl sulfoxide in dimethyl sulfide, dried, and exposed toRP Royal X-omat film (Kodak) at -70°. Interpretation of label

ing could be made by superimposition of the developed fluo-rogram on the Coomassie blue-stained protein bands.

Assay for Enzyme Activities

All assays were performed on attached monolayer cell lines(38). The determinations of enzyme activity are expressed inrelation to mg protein. Cells of semiconfluent monolayers werewashed 3 times with 2-ml aliquots of 0.9% NaCI solution,drained, and exposed to 0.6 ml of glass-distilled water. Theculture plates with lysed cells were frozen at —¿�24°.

Acid Phosphatase. After thawing, buffer (0.4 M sodiumacetate, pH 5, 0.3 ml) and substrate (0.2 M ß-glycerophos-

phate, pH 5, 0.3 ml) (Fisher Scientific Company, Fairlawn, N.J.) were added, and the plates were incubated at 37°for 2 hr.

The reaction was terminated by the addition of 0.4 ml of cold20% trichloroacetic acid. After 15 min in the cold trichloroace-

tic acid, the solution was centrifuged, and an aliquot of thesupernatant fluid (0.5 ml) was analyzed for P¡release accordingto the method of Ames and Dubin (1 ). A correction for free P¡in cells and for nonenzymic substrate hydrolysis was included.

5'-Nucleotidase. Essentially the same procedure as adopted

for acid phosphatase was used. Buffer and substrate consistedof 0.5 ml of 0.2 M Tris-HCI, pH 8.5, containing 20 mM MgCI2

and 20 mM AMP.Protein Determination. Triplicate cell cultures treated and

washed identically to the method used for cultures in thecorresponding assays were dissolved in 0.1 N NaOH, and theirprotein content was measured according to the method ofLowry era/. (31).

RESULTS

In Vivo Lung Colonization Properties

B16-F1, B16-F10, and B16-F10Lr cell lines were grown in

vitro from frozen stocks. The lung colonization capacity of the3 cell lines was tested several times over the period duringwhich the various in vitro studies were carried out. Data fromrepresentative experiments show that the relative lung colonization capacities of the 3 cell lines remained constant for 12weeks of continuous in vitro cultivation (Table 1). In all tests,cells from B16-F1 or B16-F10Lr formed fewer pulmonary mé

tastases than did equal numbers of injected B16-F10 cells.Moreover, cells from B16-F1 also formed extrapulmonary mé

tastases in organs such as liver, lymph nodes, ovary, andadrenal gland. At 18 days after i.v. injection, no extrapulmonarymétastaseswere found in mice given injections of cells of B16-F10orB16-F10Lr.

These findings are in agreement with data demonstratingthat the low lung colonization of B16-F1 and high lung colonization capacities of B16-F10 cells were maintained in cell lineseven after 18 weeks of continuous incubation in vitro.'1 Simi

larly, cells obtained from B16-F1 or B16-F10 serially trans

planted s.c. for 12 months (35 s.c. transplants) retained theirrespective low or high lung colonization capacities.5 Collec

tively then, we are confident that in vitro studies of tumorsurface properties, which may correlate with lung colonizationin vivo, can be carried out on these cell lines incubated in vitrofor up to 12 consecutive weeks. Nonetheless, we maintainedthe cell lines for no longer than 60 culture days before replacingthe cultures from frozen stocks.

Labeling of Galactose

Labeling of glycoprotein galactose residues was carried outby borohydride reduction, before or after treatment of cellswith neuraminidase to remove sialic acid and oxidation withgalactose oxidase (Fig. 1). Only traces of labeled NaB3H4

incorporation could be detected in the absence of neuraminidase treatment in all 3 cell lines. The glycoprotein profile onthe gel reveals that less radioactivity was incorporated into the

Table 1

Incidence of pulmonary métastasesin C57BL/6 mice following the i.v. injectionof B16 melanoma cells3

CelllineB16-F1B16-F10B16-F10LrCellno.612,500

25,00050,00012.500

25,00050,00050,000Median0

of pulmonary mé

tastases2(1-1 0)0

3(2-13)12(1-79)12(7-18)

100(59-145)170(125-204)2

(0-7)a

Ten mice/group.

Cells were harvested for i.v. injection from a semiconfluent culture with 2mM EDTA.

c Pulmonary métastasescounted with a dissecting microscope 18 days after

i.v. tumor cell injection.Numbers in parentheses, range.

F10

+FlOLr

4 G. Poste, personal communication.5 l. Witz and M. Baniyash, personal communication.

Fig. 1. i-iuorography pattern of sodium dodecyl sulfate-polyacrylamide gel

electrophoresis of B16 melanoma variants labeled in galactose residues of cellsurface glycoproteins. +, after neuraminidase and galactose oxidase treatment;-, galactose oxidase treatment without neuraminidase.

MAY 1980 1647

on May 21, 2020. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

A. Paz et al.

B16-F10 major glycoprotein than into the B16-F1 and F16-F10Lr major glycoprotein. The decrease in labeling is best seen

in the direct labeling of sialic acid.

Labeling of Sialic Acid

The selective labeling of cell surface sialoglycoproteinsachieved by mild periodate oxidation of the sialic acid andreduction with [3H]borohydride is illustrated in Chart 1. Densi-

tometric tracings of the gels show no qualitative differences inthe pattern of surface labeling among the 3 B16 melanoma celllines. On the other hand, the degree of labeling of the majorsialoglycoprotein (78,000 daltons) appeared to be inverselyproportional to the degree of lung colonization in vivo, high atB16-F10Lr and low at B16-F10. Similarly, the B16-F10Lr

78,000-dalton glycoprotein is more intensely labeled than thatof the B16-F10, following the galactose oxidase method of

labeling.

Lectin Binding

The composition of carbohydrate-containing receptors that

can specifically bind lectins could serve as an indicator of thesimilarity of cell surface components. The binding of radioac-

tively labeled Con A, WGA, SBA, and WBA to the surface ofB16-F, B16-F10, and B16-F10Lr is illustrated in Chart 2. The

number of Con A molecules bound to the surface of B16-F1and B16-F10 cells is identical. In B16-F10Lr cells, 50% fewer

Con A molecules are bound. A different pattern of binding isobserved with WGA. The high-lung-colonizing cell line B16-F10 had about 40% fewer WGA-binding sites than did the 2low metastatic cell lines B16-F1 and B16-F10Lr. Free SBA-

binding sites on the cell surface were barely detectable, but a30-min treatment with neuraminidase exposed the receptorsites (Chart 2). Here again, the B16-F10 cells bound 2.5-foldfewer SBA molecules than did the cells of B16-F1 or B16-F10Lr. The number of WBA surface-bound molecules revealed

that the 3 cell lines had the same number of binding sites. Nobinding of UEA1 could be detected with the 3 cell lines,

suggesting that the B16 melanoma has no exposed L-fucose

with which the lectin can interact.

Enzyme Activities

The 3 cell lines were analyzed for the specific activity of theplasma membrane enzyme marker, 5'-nucleotidase, and for

the specific activity of the lysosomal enzyme marker, acidphosphatase. The results in Table 2 show a significant reduction in the specific activity of 5'-nucleotidase in the B16-F10cell line as compared with B16-F1 and B16-F10Lr cell lines.The decrease in 5'-nucleotidase activity was not associated

with a similar decrease in cytoplasmic enzyme activities, sinceacid phosphatase activity was similar in the 3 cell lines (Table2).

Gangliosides

Gangliosides were extracted from lyophilized cell pellets andidentified as described previously (9, 45). The 3 melanoma cellvariants exhibited a very simple ganglioside pattern, which wasalmost identical to that observed when the cells were separatedon thin-layer chromatography plates. They exhibited a veryprominent GM3 and only minor traces of GM? and GMi (notshown).

Fluorescence Measurements

Fluorescence Polarization. The 3 cell lines were harvestedfrom confluent and semiconfluent conditions and were assayedfor membrane fluidity of lipid regions by means of fluorescencepolarization analysis of DPH embedded in membranes of intactviable cells (25, 41). No differences were detected in thedegree of fluorescence polarization of the 3 DPH-labeled cell

lines, grown to the same cell densities.Fluorescence Photobleaching Recovery. Fluorescence

photobleaching measurements (39) of diffusion of the lipidprobes DM and rhodamine-conjugated WGA on the plasmamembrane of single cells of B16-F10 and B16-F10Lr demon

strated no differences in either diffusion constants or mobile

i i i 111 i i i i i200 150 100 70 50 30 20 FRONT 200 150 100 70 50 30 20 FRONT 200 150 100 70 50 30 20 FRONT

MOLECULAR WEIGHT ( x 10'3)

Chart 1. Densitometric tracing of a fluorogram of sodium dodecyl sulfate-polyacrylamide gel electrophoresis of B16 melanoma variant cell lines. Glycoproteinswere labeled in the terminal sialic acid. A, B16-F1; B, B16-F10; C, B16-F10U.

1648 CANCER RESEARCH VOL. 40

on May 21, 2020. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Surface Properties of Metastatic Cells

24

-- 20x

0.

Ãœ 16•¿�o

12

ion

A 60•¿�o

m

g 40S

CM20

10 20 30 40 50 60 70 80 90 100125l-Con A(Mg/ml)

10 20 30 40125|-WGA l^g/ml)

50

_ 1201b

«IO

Q.

.^ 4

10 20 30 70 80 10040 50 60.. . 125l-WBA(Mg/ml)

Chart 2. Specific binding of iodinated lectins to B16-F1 (O, •¿�),B16-F10 (A, A), and B16-F10Lr (O, •¿�)in I25I-SBA binding curve. O, A, D, cells pretreated with

äuraminidase;•¿�,A. •¿�without neuraminidase treatment. Experiments were done in quadruplicate. Bars, S.D.

Table 2Specific activities ol 5'-nuc/eotidase and acid phosphatase in B16 melanoma

cell lines

CellsB16-F1

B16-F10B16-F10Lr5'-Nucleotidase

activity(nmol P,/mgprotein)142.2

±2.1a90.9

±1.5176 ±3.5Acid

phosphatase activity(nmol P,/mgprotein)1066.6

±15.9919.5 ± 8.1

1046 ±11.51Mean ±S.E. computed from 6 different samples.

fractions within the 20 to 30% S.D. of the measurements. Thediffusion of the DMmolecules in the membrane plane was ~150x 1010 sq cm/sec, while the diffusion of rhodamine-WGA was~3 x 10'°sq cm/sec. These values were similar in both cell

lines.

Morphological Observation

Treatment of the cells with 2 HIM EDTA to prepare cells fori.v. injection induced cell rounding and detachment from thesubstratum. The surface of the detached cells is rough, andthe cells have a moderate number of flaps and ridges withextended thin filopodia. The cells of the B16-F10Lr cell line

appeared to be somewhat smaller than the cells of the B16-F1or B16-F10 lines. Thus, the differences in cell arrest in the lungare not due to gross variations in the morphology or size amongthe 3 B16 melanoma cell lines.

DISCUSSION

In this study, we have examined some aspects of cell surfacecomposition and properties of 3 B1 6 melanoma variant celllines with differing capacity to form experimental pulmonarymétastasesfollowing i.v. injection. The incidence of pulmonarymetastasis produced by cells of B16-F10 line was at least 10-fold that of B16-F1 or B16-F10Lr (Table 1). This difference in

biological behavior in vivo was not reflected by comparablemajor qualitative alterations in the cell surface characteristicsthat we studied.

The process of metastasis is complicated, and the productionof visible métastasesis dependent on the ability of malignantcells to survive and complete a number of sequential steps.The injection of tumor cells into the circulation bypasses thefirst steps of the process, i.e., invasion and detachment. However, all the subsequent steps such as transport and survival inthe circulation, arrest in the capillary bed of an organ, extravasation into the parenchyma, establishment of a microenvi-

ronment, escape from host defense mechanisms, and tumorcell multiplication in the distant organ must all be completedfor clinical metastasis to occur. Failure to complete any one ofthese steps will lead to the elimination of a tumor cell as apotential progenitor of a secondary growth (42).

Reports from several laboratories have now demonstratedthat the in vivo metastatic capacity of the B16-F lines correlates

MAY 1980 1649

on May 21, 2020. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

A. Raz et al.

with their ability to aggregate with platelets (20), a phenomenonthat could be due to differences in prostaglandin production(1 7). The cell lines also differ in their ability to clump withsyngeneic lymphocytes (13) and endothelial cells (35, 48) andin their response to hormones (36). Control of many of theseinteractions probably resides at the tumor cell surface, and itis not unreasonable to postulate that differences in cell surfacecomposition may account for the disparate behavior of thevariant cell lines. In fact, some isolated cell surface differencesbetween the cell lines have been described (4, 50), but nocorrelation between these differences and divergent metastaticbehavior has been possible. Since no one property of tumorcells may determine the eventual outcome of metastasis, weattempted to survey a wide range of tumor cell surface characteristics and then compared the overall profile of one cell linewith the others. We wish to emphasize 2 points regarding theexperimental procedures: (a) this study represents a collaborative effort of 5 laboratories, in which each concentrated onassays that they use routinely and with which they are mostfamiliar; (b) the cell lines were coded for all in vitro assays, andthus the studies represent blind assays.

We have used several different techniques of labeling surfaceconstituents since most techniques, if used in isolation, havesome drawbacks (8). Protein composition of the cell surfacewas studied by protein tyrosine iodination with lactoperoxidase(32). Analysis of surface-labeled proteins failed to demonstrateany appreciable difference in the surface-labeling pattern ofthe 3 B16-F lines (data not shown). These results, which failedto demonstrate any difference between B16-F1 and B16-F10cell lines, are in agreement with an earlier report (34).

Studies of cell surface carbohydrates used 3 techniques: (a)labeling of galactose residues; (b) surface labeling of sialic acidresidues; and (c) lectin binding. Analysis of cell surface gly-coproteins by gel electrophoresis and fluorography prior to andafter neuraminidase treatment demonstrated that most of theterminal galactose residues in the carbohydrate chains of gly-

coproteins are sialated, since no appreciable labeling wasdetected prior to the cleavage of the sialic acid. At least 11distinct bands of surface glycoproteins ranging from 14,000 to200,000 daltons, with a major glycoprotein at 78,000 dallons,could be demonstrated. Direct labeling of the sialic acid andanalysis of the labeled sialoglycoprotein by gel electrophoresisgave a pattern similar to the one obtained with the galactoseoxidase method. Both techniques failed to demonstrate anyqualitative differences among the B16-F1, B16-F10, and B16-F10Lr cell lines.

Lectin binding of the cell lines differed. B1 6-F1 and B16-F10

cells exhibited similar binding of Con A. This is in agreementwith studies reporting similar agglutination of these 2 cell linesby Con A (34). However, the poorly metastatic cell line, B16-F10Lr, had a significantly lower number of surface Con Areceptors than the 2 other lines. The B16-F10Lr cells are

resistant to in vitro lysis mediated by syngeneic lymphocytes,which is presumably due to the deficiency of lymphocyte-

binding receptors (13). Whether the receptors for Con A andcytotoxic syngeneic lymphocytes are related is unclear. TheB16-F10 cells were found to have about a 50% reduction inWGA or SBA binding as compared with B16-F1 and B16-F10Lr

cells. SBA agglutinin, which binds to terminal galactose residues, barely bound to the surface of the 3 cell lines prior to theremoval of sialic acid. This finding correlated with the results

that we obtained by surface labeling with galactose oxidase.Moreover, the fact that after neuraminidase treatment fewerSBA-binding sites were exposed on the F10 cell membrane

than on the other 2 cell lines also agrees with the data obtainedby the galactose oxidase labeling. The latter technique demonstrated that surface galactose accessible for labeling wasreduced on B16-F10 cells and that B16-F10 cells also have

less surface sialic acid accessible for labeling. We did not findthat B16-F10 cells have a 2-fold increase in neuraminidase-

accessible sialic acid surface molecules (4, 50). The 3 celllines had the same number of WBA-binding sites, reflecting the

fact that the sugar residues, which are in proximity to theprotein core, are the same in all 3 cell lines.

Plasma membrane gangliosides are believed to be receptorsfor numerous biologically active agents such as toxins andhormones. Moreover, changes in ganglioside pattern may accompany transformation (21). We found no qualitative differences in the ganglioside composition of the 3 cell lines. Theganglioside pattern was found to be a simple one, consistingmainly of GM3with traces of GM2and GMi. These findings agreewith a previous report in which the ganglioside profile of B16-F1 and B16-F10 cells was studied (50).

We next examined the relationship of protein dynamics tomembrane lipid fluidity. Fluorescence polarization analysis ofDPH when embedded in membrane lipids was used to monitorthe degree of microviscosity of the surface membrane of B16-F10, B16-F1, and B16-F10Lr tumor cells. We detected no

differences in lipid microviscosity among cells of the 3 variantlines. The fluorescence polarization analysis, which examinespopulations of cells, is an averaging method. In contrast, fluorescence photobleaching recovery measures the diffusion ofthe lipid probe DM and rhodamine-conjugated WGA in the

plasma membrane of individual cells. The measurements wereperformed at a given lectin concentration on the same day.Nonetheless, no detectable differences in either diffusion constant or mobile fractions were observed between cells of B16-F10 and B16-F10U lines.

The adenosine-generating enzyme 5'-nucleotidase is a

plasma membrane-bound enzyme of mammalian cells. Differentlevels of this enzyme can distinguish among different subpop-ulations of cells both in normal and transformed systems (2,30). The specific activity of 5'-nucleotidase in the highly met

astatic melanoma cell line (B16-F10) was decreased by about50% as compared with the 2 low-metastatic cell lines (B16-F1and B16-F10Lr). The observed reduction of enzyme activity

was limited to the plasma membrane, since it was not accompanied by a similar reduction of an intracellular enzyme, acidphosphatase. The B16 melanoma variants had a relatively lowspecific activity of 5'-nucleotidase as has been observed in

other transformed cells (37).Collectively then, the present data as well as that of others

(34, 48) are unable to demonstrate a major alteration in theplasma membrane characteristics of B16 melanoma cells thatmight be associated with their metastatic potential in vivo. Theproduction of a metastasis requires a cell to survive manydestructive events. It is entirely possible that what appears tobe a subtle variation in a cell surface property measured invitro could be responsible for a dramatic alteration in malignantbehavior in vivo. Certainly, our study illustrates some of theproblems inherent in evaluating a complex in vivo process suchas metastasis by using in vitro correlates which may measure

1650 CANCER RESEARCH VOL. 40

on May 21, 2020. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Surface Properties of Metastatic Cells

only one cellular property (of many) requisite for the metastaticcell. Nonetheless, it is possible that subtle variations in composition (47) or arrangement of cell surface constituents couldindeed be responsible for the ability of tumor cells to interactwith their host, survive, and develop into métastases.

REFERENCES

1. Ames, B. N., and Dubin, D. T. The role of polyamines in the neutralization ofbacteriophage deoxyribonucleic acid. J. Biol. Chem.. 235. 769-771, 1960.

2. Barton. R. W., and Goldschneider, I. 5'-Nucleotidase activity in subpopula-tion of rat lymphocytes. J. Immunol.. 121: 2329-2334, 1978.

3. Bonner, W. M., and Laskey, R. A. A film detection method for tritium-labeledproteins and nucleic acids in polyacrylamide gels. Eur. J. Biochem.. 46: 83-

88, 1974.4. Bosmann, H. B., Bieber, G. F., Brown, A. E., Case, H. R., Gersten, D. M.,

Kimmerer, T. W., and Lione, A. Biochemical parameters correlated withtumor cell implantation. Nature (Lond.), 246: 487-489, 1973.

5. Briles, E. B., and Kronfeld, S. Isolation and metastatic properties of detachment variants of B16 melanoma cells. J. Nati. Cancer Inst.. 60. 1217-1222.1978.

6. Brunson, K. W., Beattie, G., and Nicolson, G. L. Selection and alteredproperties of brain-colonizing metastatic melanoma. Nature (Lond.), 272.543-545, 1978.

7. Brunson. K. W., and Nicolson, G. L. Selection and biologic properties ofmalignant variants of a murine lymphosarcoma. J. Nati. Cancer Inst., 67:1499-1503. 1978.

8. Carraway, K. L. Covalent labeling of membranes. Biochim. Biophys. Acta.415: 379-410, 1975.

9. Dem, H., Sela, B-A., Roseman, S., and Sachs, L. Block in gangliosidesynthesis in transformed hamster cells and their revenants. J. Biol. Chem.,249. 659-661. 1974.

10. Dexter. L. D., Kowalski, H. M., Blazar, B. A., Fligiel, A., Nogel. R., andHeppner, G. H. Heterogeneity of tumor cells from a single mouse mammarytumor. Cancer Res.. 38. 3179-3181. 1978.

11. Fidler, I. J. Metastasis: quantitative analysis of distribution and fate of tumoremboli labeled with 125l-5-iodo-2'-deoxyuridine. J. Nati. Cancer Inst., 45:

775-782, 1970.

12. Fidler, I. J. Selection of successive tumor lines for metastasis. Nat. NewBiol., 242: 148-149, 1973.

13. Fidler, I. J., and Bucana. C. Mechanism of tumor cell resistance to lysis bysyngeneic lymphocytes. Cancer Res., 37: 3945-3956, 1977.

14. Fidler, l. J., Gersten, D. M., and Budmen. M. B. Characterization in vivo andin vitro of tumor cells selected for resistance to syngeneic lymphocyte-mediated cytotoxicity. Cancer Res., 36: 3160-3165, 1976

15. Fidler, l. J., Gersten, D. M., and Hart, I. R. The biology of cancer invasionand metastasis. Adv. Cancer Res., 28: 149-250, 1978.

16. Fidler, I. J., and Kripke, M. L. Metastasis results from preexisting variantcells within a malignant tumor. Science (Wash. D. C.), Õ97:893-895. 1977.

17. Fitzpatrick, F. A., and Stringfellow, D. A. Prostaglandin D2 formation bymalignant melanoma cells correlates inversely with cellular metastatic potential. Proc. Nati. Acad. Sei. U. S. A., 76: 1765-1769, 1979.

18. Fogel, M., Gorelik, E., Segal, S., and Feldman, M. Differences in cell surfaceantigens of tumor métastasesand those of the local tumor. J. Nati. CancerInst., 62: 585-588. 1979.

19. Gahmberg. C. G.. and Anderson. L. C. Selective radioactivity labeling of cellsurface sialoglycoproteins by periodate-tritiated borohydride. J. Biol. Chem.,252: 5888-5894, 1977.

20. Gasic, G. J., Gasic, R. B., Galanti, N., Johnson, T., and Murphy. S. Platelet-tumor cell interaction in mice. The role of platelets in the spread of malignantdisease. Int. J. Cancer, 11: 704-718, 1973.

21. Hakomori. S. Structures and organization of cell surface glycolipids dependency on cell growth and malignant transformation. Biochim. Biophys. Acta,41 7: 55-89, 1975.

22. Hayat, M. A. Positive Staining for Electron Microscopy, pp. 162-176. New

York: Van Nostrand Reinhold Co., 1975.23. Hunter, W. M. The preparation of radioiodinated proteins of high activity,

their reaction with antibody in vitro, the radioimmunoassay. in: D. M. Weir(ed.). Handbook of Experimental Immunology, pp. 808-842. Philadelphia:

F. A. Davis Co., 1967.24. Hynes, R. O. Cell surface proteins and malignant transformation. Biochim.

Biophys. Acta, 458: 73-107, 1976.25. Inbar, M., Yuli, I., and Raz, A. Contact-mediated changes in the fluidity of

membrane lipids in normal and malignant transformed mammalian fibro-

blasts. Exp. Cell Res., 105: 325-335, 1977.

26. Kerbel, R. S.. Twiddy, R. R., and Robertson, D. M. Induction of a tumor withgreatly increased metastatic growth potential by injection of cells from a low-metastatic H-2 heterozygous tumor cell line into an H-2 incompatible parental strain. Int. J. Cancer, 22. 583-594. 1978.

27. Killion, J. J., and Kollmorgen. G. M. Isolation of immunogenic tumor cells bycell-affinity chromatography. Nature (Lond.). 259: 674-676, 1976.

28. Kripke, M. L., Gruys, E., and Fidler, I. J. Metastatic heterogeneity of cellsfrom an ultraviolet light-induced murine fibrosarcoma of recent origin. Cancer Res., 38 2962-2967, 1978.

29. Laemmli, U. K. Cleavage of structural proteins during assembly of the headof bacteriophage T4 Nature (Lond.). 227: 680-685. 1970.

30. Lopez, J., Zucker-Franklin, D., and Silber, R. Heterogeneity of 5'-nucleotid-

ase in lymphocytes in chronic lymphocytic leukemia. J. Clin. Invest.. 52:1297-1300. 1973.

31. Lowry, O. H., Rosebrough. N. J.. Farr, A. L.. and Randall. R. J. Proteinmeasurement with the Polin phenol reagent. J. Biol. Chem., /93: 265-275,

1951.32. Marchalonis. J. J., Cone, R. E., and Sanier. V. Enzymic iodination: a probe

for accessible surface proteins of normal and neoplastic lymphocytes.Biochem. J., Õ24.921-927, 1971.

33. Nicolson. G. L. Trans-membrane control of the receptors on normal andtumor cells. II. Surface changes associated with transformation and malignancy. Biochim. Biophys. Acta. 458: 1-72. 1976.

34. Nicolson, G. L., Birdwell, C. R.. Brunson, K. W., Robbins, J C., Beattie, G.,and Fidler, I. J. Cell interaction in the metastatic process: some cell surfaceproperties associated with successful blood-borne tumor spread in cell andtissue interaction. In: J. W. Lash and M. M. Burger (eds ). Cell and TissueInteraction, pp. 225-241. New York: Raven Press, 1977.

35. Nicolson. G. L , and Winkelhake, J. L. Organ specificity of blood-bornemetastasis determined by cell adhesion. Nature (Lond.), 255: 230-232.

1975.36. Niles, R. M., and Makarski, J. S. Control of melanogenesis in mouse

melanoma cells of varying metastatic potential. J. Nati Cancer Inst., 6f:523-526, 1978.

37. Raz, A., Collard. J. G., and Inbar, M. Decrease in 5'-nucleotidase activity in

malignant, transformed, and normal stimulated cells. Cancer Res.. 38:1258-1262, 1978.

38. Raz, A.. Shahar, A., and Goldman, R. Characterization of an in vivo inducedperitoneal macrophage population following intraperitoneal injection of con-canavalin A. RES J. Reticuloendothel. Soc., 22: 445-460, 1977.

39. Schlessinger, J., Axelrod, D., Koppel, D. E., Webb, W. W., and Elson. E. C.Lateral transport of a lipid probe and labeled proteins on a cell membrane.Science (Wash. D. C.). ÃŽ95.307-309, 1977.

40. Schwartz, S. M. Selection and characterization of bovin aortic endothelialcells. In Vitro, 14: 966-980, 1978.

41. Shinitzky. M., and Inbar, M. Difference in microviscosity induced by differentcholesterol levels in surface membrane lipid bilayer of normal lymphocytesand malignant lymphoma cells. J. Mol. Biol.. 85: 603-615, 1974.

42. Sugarbaker. E. V. Cancer metastasis: a product of tumor-host interactions.Current Problems in Cancer, 3 1-59, 1979.

43. Suzuki. N., and Withers, H. R. Isolation from murine fibrosarcoma of celllines with enhanced plating efficency in vitro. J. Nati. Cancer Inst., 60: 179-

183, 1978.44. Suzuki. N., Withers, H. R.. and Koehler. M. W. Heterogeneity and variability

of artificial lung colony-forming ability among clones from mouse fibrosarcoma. Cancer Res., 38: 3349-3351, 1978.

45. Svennerholm. L. Quantitation estimation of sialic acids. II. A colorimetrieresorcinol-hydrochloric acid method. Biochim. Biophys. Acta. 24: 604-611,

1957.46. Tao, T. W.. and Burger, M. M. Nonmetastasizing variants selected from

metastasizing melanoma cells. Nature (Lond.), 270: 437-438, 1977.47. Warren. L.. Buck, C. A., and Tuszynski, G. P. Glycopeptide changes and

malignant transformation: a possible role for carbohydrate malignant behavior. Biochim. Biophys. Acta, 516: 97-127, 1978.

48. Warren, L., Zeidman, I., and Buck, B. A. The surface glycoproteins of amouse melanoma growing in culture and as a solid tumor in vivo. CancerRes., 35: 2186-2190. 1975.

49. Winkelhake. J. L.. and Nicolson, G. L. Determination of adhesive propertiesof variant metastatic melanoma cells to BALB/3R3 cells and their virustransformed derivatives by a monolayer assay. J. Nati Cancer Inst . 46:285-298, 1976

50. Yogeeswaran, G.. Stein, S. B., and Sebastian, H. Altered cell surfaceorganization of gangliosides and sialyglycoproteins of mouse metastaticmelanoma variant lines selected in vivo for enhanced lung implantation.Cancer Res., 38: 1336-1344, 1978.

MAY 1980 1651

on May 21, 2020. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1980;40:1645-1651. Cancer Res   A. Raz, W. L. McLellan, I. R. Hart, et al.   Metastatic PotentialCell Surface Properties of B16 Melanoma Variants with Differing

  Updated version

  http://cancerres.aacrjournals.org/content/40/5/1645

Access the most recent version of this article at:

   

   

   

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

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/40/5/1645To request permission to re-use all or part of this article, use this link

on May 21, 2020. © 1980 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from