[CANCER RESEARCH 44, 2668-2676, June 1984]

Cholesterol-induced Growth Stimulation, Cell Aggregation, and MembraneProperties of Ascites Tumor Cells in Culture1

E. W. Haeffner,2 C. J. K. Hoffmann, M. Stoehr, and H. Scherf

Institut fürZell-und Tumorbiologie [E. W. H., C. J. K. H.]; Institut fürexperimentelle Pathologie [M. S.]: und Institut fürToxikologie und Chemotherapie [H. S.], DeutschesKrebsforschungszentrum. 0-6900 Heidelberg, Federal Republic of Germany

ABSTRACT

Ascites tumor cells can be cultivated at a reduced serumconcentration if cholesterol (2.50 mg per 100 ml of medium) isadded to the culture medium. At serum concentrations of 3%,optimal growth properties are obtained; below 3%, cell culturesusually perish after a few days. Cells grown in the presence ofadded cholesterol have an elevated content of this molecule percell as well as in the plasma membrane, and they also show acholesterol concentration-dependent rate of proliferation. Precursors of the cholesterol-biosynthetic pathway like mevalonicacid, added in mw amounts, or squalene and lanosterol cannotbe substituted for cholesterol itself. This is due to the observationthat the biosynthetic pathway is blocked at the stage of lanosterol conversion to cholesterol. Cholesterol de novo synthesisfrom acetate is regulated by the cholesterol content of the cells,which also affects the production of ubiquinone and dolichol.Growth factors such as insulin, prostaglandin F2o,and transferrinadded to the medium do not mimic the cholesterol-induced effect.Distribution of DNA during cell cycle and the cell density-depend

ent reduction in macromolecule synthesis is very similar to thecontrol cells. In contrast, cells without added cholesterol showreduced growth properties accompanied by the accumulation ofcells in the mitotic and G2 phase.

The cholesterol/phospholipid ratio of the plasma membranesof cholesterol-rich cells is about 15% lower than of the controlcells and 40% higher compared to the cholesterol-poor cells,

which, however, does not significantly alter the membrane fluiditybetween the cholesterol-rich and -poor cells as revealed byfluorescence polarization measurements. The most dramaticbehavior of the cholesterol-rich cells is their tendency to formaggregates, which is demonstrated either by concanavalin A-induced agglutination or by cell density-dependent aggregation

shown by interference microscopy in vivo.

INTRODUCTION

The requirement of some cell types for preformed lipids,especially cholesterol, normally supplied by the serum has recently been established (27, 29, 32, 33). Brennemann ef al. (9)have shown that ascites cells take up major amounts of cholesterol from the ascites fluid and, in the same study, it has furtherbeen found that the cholesterol de novo synthesis is rather low.In another paper (26), we have reported that ascites tumor cells

1This work was supported by funds of the federal government, Federal Republicof Germany, and by funds of the state government of Baden-Württemberg.

2To whom requests for reprints should be addressed, at Institut fürZell-undTumorbiologie, German Cancer Research Centre, P. 0. Box 101 949, D-6900

Heidelberg, Federal Republic of Germany.Received September 21,1982; accepted March 12,1984.

do not grow when cultivated in delipidated serum medium.Attempts to induce cell proliferation in delipidated serum mediumto which cholesterol has been added were only partially successful.3 In this study, we have investigated the effect of a

reduced serum concentration with and without the addition ofcholesterol upon ascites cell growth, lipid de novo synthesis, andsurface properties. We present evidence for permanent culturesof these cells at drastically reduced serum concentrations, andwe also show data revealing changes at the surface which maybe responsible for inducing a highly increased cell-cell interaction.

Preliminary results of these experiments have been reportedelsewhere (25).

MATERIALS AND METHODS

Materials. Concanavalin A, linoleic acid, palmitoleic acid, linolelaidicacid, prostaglandins E and F2„,phenylmethylsulfonyl fluoride, 4,6-diami-dino-2-phenylindol, and 1,6-diphenyl-1,3,5-hexatrien were purchased

from Serva, Heidelberg, Federal Republic of Germany, and insulin plustransferrin was from Sigma, Munich, Federal Republic of Germany. Theradioisotopes [mef/>y/-3H]thymidine, 2-[14C]acetic acid, and N-[acefy/-3H]

concanavalin A were purchased from New England Nuclear, Dreieichen-

hain, Federal Republic of Germany.Cell Cultures. Hyperdiploid ascites tumor cells, type Karzel, were

cultivated in minimum essential medium with Earle's salt, 1% 4-(2-

hydroxyethyl)-1 -piperazineethanesulfonic acid buffer, pH 7.4, and horseserum from Boehringer-Mannheim, Federal Republic of Germany. The

serum concentration in the medium varied between 0.5 and 10% asindicated in the chart legends. Cell cultures were seeded at an initial celldensity of about 10s cells/ml using sterile 75-sq cm Corning tissue culture

flasks. The medium was changed every second day by centrifugation ofthe cells, and cell densities were determined with the aid of a Neubauerhemocytometer. Viability not lower than about 98% was checked by thetrypan blue exclusion test. Additions of cholesterol to the serum wereperformed similar to the procedure of Corwin ef al. (15). Between 0.4and 2.5 mg of cholesterol dissolved in 12.5 to 75 /il each of isopropylalcohol and propylene glycol were mixed with 0.4 to 2.5 ^l of an ethanolicsolution of 42 to 250 ^g of linoleic or any other fatty acid neutralizedwith 1 N NaOH. This mixture was carefully added to 10 ml of serum insuch a way that no precipitation occurred. The clear serum was thencombined with 90 ml of minimum essential medium with Earle's salt and

1% 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid buffer, pH 7.4. Agrowth-stimulating effect of linoleic acid was excluded: (a) by reducingthe fatty acid concentration to one-tenth of the amount normally used;

and (o) by replacing linoleic acid through other acids like oleic, palmitoleic,or linolelaidic acid. Under both conditions, similar growth properties ofthe ascites cells were observed as with linoleic acid itself. Also, propyleneglycol and isopropyl alcohol had no effect on cell growth.

Plasma Membrane Preparation. Cells were disrupted after swellingin 2 ITIMEDTA, pH 7.4, in hypotonie salt solution according to the method

3C. J. K. Hoffmann, N. Pawetetz, R. Friedel, and E. W. Haeffner, unpublished

observations.

2668 CANCER RESEARCH VOL. 44

on March 5, 2020. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Control of Growth and Surface Properties ofAscites Cells

of Marnarli ef al. (31). A nuclear-free 12,000 x g pellet was prepared

from the cell homogenate by differential centrifugation which was thenused as starting material for the purification of the plasma membranesby gel filtration on Sephacryl S-1000 superfine (Pharmacia Fine Chemi

cals, Uppsala, Sweden). The membranes were eluted with 0.25 M sucrose, 10 mM Tris-HCI, pH 7.4, at a flow rate of 50 ml/hr. After collectionof the plasma membrane-containing fractions, the material was dialyzedagainst 10 mw Tris-HCI, pH 7.5, containing 0.2 mw MgCI;, and phenyl-methylsulfonyl fluoride as protease inhibitor, was freeze-dried, and theresidual material was taken up in Tris-buffered isotonic sucrose solution

for further experiments. Details of the enzymatic assays have beendescribed elsewhere (20).

Labeling Experiments. Thymidine incorporation rates were determined by using [mef/7y/-3H]thymidine (specific activity, 20 Ci/mmol) at

concentrations of 0.25 to 0.75 /¿Cito 0.5-ml aliquots of the cell suspensions (2.5 x 103 to 7 x 104 cells) and incubated at 37° for 30 min.Determination of labeled cholesterol derived from [14C]acetate was asfollows. Amounts of 5 x 106 cells were taken from the culture flasks;

centrifuged; washed once with phosphate-buffered saline, pH 7.4, con

sisting of 0.138 M NaCI, 2.7 mw KCI, 0.49 mM MgCI2, 0.90 mM CaCI2,1.47 mw KH2PO4, and 8.1 mw Na2HPO4; and suspended in 2 ml of thesame buffer, to which 62.5 ¿¿Ciof [2-'4C]acetate (specific activity, 51MCi/Mtnol) were added. After incubation of the reaction mixtures at 37°

for the time indicated in the chart, 0.6-ml aliquots were taken andcentrifuged, and the pellets were washed 3 times with phosphate-

buffered saline, pH 7.4. The cell lipids were then extracted by the Folchprocedure (18) and analyzed by TLC4 using Sil G-covered plastic sheets

(Macherey-Nagel, Duren, Federal Republic of Germany) and petroleum

ethendiethyl etheracetic acid (80:20:2) as solvent system. After visualization with iodine vapor and disappearance of the color, the spots werecut off the plates and quantitated by liquid scintillation counting.

For the determination of cholesterol precursors and polyisoprenoids,the Folch extracts were saponified with 1 ml of 5% NaOH in methanolfor 1 hr at 100° in a sealed ampul. The unsaponifiable material wasextracted 3 times with petroleum ether (b.p. 40-60°) and then analyzed

by TLC with chloroform as solvent system. Radioactivity distribution inthe various spots was performed as described above.

Quantitäten of cholesterol was done by gas-liquid chromatography

(21), and for comparison by an enzymatic procedure according to themethod of Röschlau (35). Pi was determined by the method of Eibl andLands (16). The membrane polar lipids were quantified by phosphorusanalysis after TLC separation using Sil G-covered plastic sheets. Micro

scopic examinations of the cell suspensions were made by carefullyputting the cells on glass plates and analyzing them by interferencemicroscopy.

DNA Histograms. Cells were washed with 0.85% NaCI solution(isotonic salt solution), fixed with 70% ethanol, and collected at thisstage at 4°.After removal of ethanol, the above buffer containing the

fluorochrome 4,6-diamidino-2-phenylindol (3 Mg/ml) was then added.

Flow cytometric analysis was carried out with a computerized FACS IIcell sorter (Becton Dickinson) which was equipped with a UV laser beam.The laser was tuned at 363 nm; cellular fluorescence was collectedabove 390 nm.

Fluorescence Polarization Measurements. Fluorescence polarization was measured at 90°angle relative to the exciting beam, using a

Perkin-Elmer MPF 4 fluorescence spectrometer. Two commercial polar

izers (Zeiss, Oberkochen, Federal Republic of Germany) were mountedat the excitation (360 nm) and emission (426 nm) site of the cuvet. Thesamples were excited with vertically polarized light, and the intensity ofthe polarized fluorescence emission was measured both at vertical andhorizontal orientation of the emission polarizer. The labeling conditionsof the membrane preparations are described in Fig. 1. The 1,6-diphenyl-

4The abbreviations used are: TLC, thin-layer chromatography: C/P, cholesterol/

phospholipid.

1,3,5-hexatrien-labeled samples were exposed to the excitation light

shortly before the measurement to eliminate the possibility of reversiblebleaching of 1,6-diphenyl-1,3,5-hexatrien (39). The temperature was

controlled with a Haake thermostatic bath, equipped with a PG40 temperature programming device and measured directly within the cuvet bymeans of a digital thermometer. Before each measurement, the membrane suspensions were mixed gently to ensure isotropie distribution.The temperature profiles were measured by scanning from the lowesttemperature to the highest. The preparation of the samples was madeaccording to the procedure of Fuchs et al. (19), and the calculation ofthe polarization, according to the formula given by Shinitzky and Inbar(40). Corrections were made for scattered light of the unlabeled cellsuspensions (1.4 to 7%).

Agglutination Assay. The test was made similar to the procedurewhich has been described by Rule ef al. (36). Amounts of 106 cells were

taken from the culture flasks and centrifuged to remove the medium.The pellets were suspended in 2 ml of phosphate-buffered saline, pH

7.4, in polystyrene tubes to which 501¿\of a 2% concanavalin A solutionwere added. The mixtures were incubated at 37° under shaking at a

rate of 170 cycles/min, and at various times, aliquots were taken forsingle cell counting using a Neubauer hemocytometer.

Concanavalin A Binding Studies. The binding experiments wereperformed essentially according to Feller ef a/. (17). Aliquots of 5 x 105cells in 1 ml of phosphate-buffered saline were incubated at 4°for 1 hrin the presence of 3 to 300 ^g of [3H]concanavalin A (0.035 to 3.5 ßd).

After extensive washing of the cells, the bound radioactivity was measured. For the determination of specific binding, the cells were countedbefore and after treatment with a-methylmannoside. Evaluation of the

data was carried out according to the method of Scatchard (38).

RESULTS

Initial studies about the relationship between serum concentration and ascites cell growth have shown that these cells canbe cultivated at reduced serum concentrations (from 10 to 5%)without any significant reduction in growth properties. However,a reduction of the serum concentration below 5% results in aretardation of cell proliferation with the consequence that thecells perish after a few days. Chart 1 shows a comparison ofcells grown under normal and reduced, i.e., 0.5 and 3% serumconcentrations. Cells with 0.5% serum in the medium do notgrow at all, whereas cells at 3% serum proliferate until about 5days. Addition of some selected well-known growth factors such

as insulin, transferrin, and prostaglandins either alone or incombination have no stimulating effect. However, when cholesterol is added in a concentration of 2.50 mg/100 ml of medium,then we observe some growth in the case of 0.5% serum andnearly optimal growth in the case of 3% serum concentration. Inthe latter case, cells can be kept as long-term cultures. Supple

mentation of the medium with linoleic acid alone, which is addedin combination with cholesterol, does not stimulate cell growthat a substantial rate, which has also been found for normal andvirus-infected chicken embryo fibroblasts (22). Furthermore, reduction of the linoleic acid concentration to one-tenth of the

amount normally added does not have any effect on cell growth.Also, the substitution of linoleic by other natural and unnaturalfatty acids like oleic, palmitoleic, or linolelaidic acid, has noinfluence on the growth properties. The growth stimulation isspecific for cholesterol. This has been investigated by replacingcholesterol with various precursors of the cholesterol-biosyn-

thetic pathway such as mevalonic acid in mw concentrations,squalene and lanosterol. Neither of these precursors can initiate

JUNE 1984 2669

on March 5, 2020. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

E. W. Haeffner et al.

150— ISO—i

100—

O 50-

B

2345

DAYS

Chart 1. Growth curves of ascites cells in culture. A, curves at 0.5% and 8, at 3% serum concentration in the medium. Control serum ( ), serum plus 2.50 mg ofcholesterol and 250 *ig of Cla:2 per 100 ml of medium (• •),the same as before but with 25 >ig C,,* (D D), serum minus cholesterol (O O), serum plusgrowth factors such as insulin, transform, or prostaglandin (V V) at concentrations between 1 and 4 ^g/10s cells.

140- 140-120-—

121x

100-

0o0

80-*

60-en_iio

40--20-

~to-;B//

///*//

^4/

/'

'/'' --'//'

/'*/

/,''/'

.-''

*---*'••'''*+

^-~"|

l | ,|2

3 4EDRYS

DRYSChart 2. Growth curves of ascites cells. A, in control medium with 10% serum (D G), in medium with 3% serum supplemented with 38.7 jimol (0.39 rriM) mevakxiic

acid (x. x), and 6.5 /imo! each of squatene (A A) and lanosterol (• «)per 100 ml of medium; B, in medium with 3% serum supplemented with increasingconcentrations of cholesterol, 0.4 mg (x x), 1.25 mg (A A), and 2.5 mg/100 ml of medium (« •).

2670 CANCER RESEARCH VOL. 44

on March 5, 2020. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Control of Growth and Surface Properties of Ascites Cells

the growth-stimulating effect of cholesterol (Chart 2A), and it hasalso been shown that the rate of proliferation is dependent uponthe cholesterol concentration in the medium (Chart 2B). Thefeedback control of cholesterol synthesis has generally beenaccepted. To test this also in case of ascites tumor cells, wehave measured the rate of cholesterol de novo synthesis from

6.5-

6.0-

cn 5.5-_l_i

O 4.5-

214.0-

3.5-

3.0-

x

CM

O 2.5HOo~ 2.0-

x•s- i.sH

0.00.0 1.5

—I

2.00.5 1.0HOURS

Chart 3. Synthesis of cholesterol from labeled acetate of cells cultivated at 3%serum concentration in the presence (x x) and absence p G) of externalcholesterol in the medium. Details of the experimental conditions are describedunder "Materials and Methods." Curves, mean of 3 individual experiments. MIO.,

million.

6605

I

aa. 'u

885

SLCUD SLCUDChart 4. Acetate incorporation into the nonsaponifiable lipid fraction. Aliquots

of 4 x 10°cells were incubated in the presence of 12.5 ^Ci of 2-[14C)acetate at37°for 2 hr. After Folch extraction and separation into the nonsaponifiable and

sapomfiable fraction, the radioactivity distribution was determined as describedunder "Materials and Methods.' A. at 3% serum without and, B, in the presence of

cholesterol; 6605 and 885, total incorporated radioactivity (cpm). S, squalene; L,lanosterol; C. cholesterol; U. ubiquinone; D. dolidol.

labeled acetate of cells grown in 3% serum medium either withor without cholesterol added to the medium. The results (Chart3) indicate that the rate of acetate incorporation is depressed bya factor of about 6 in the cholesterol-supplemented cells. Similar

results are obtained for acetate incorporation into the entiresterol fraction and the radioactivity distribution between cholesterol precursors and the products of the branched pathways,ubiquinone and dolidol, is shown in Chart 4. Unlike in liver, themajor labeled products were found to be lanosterol, followed bysqualene, and the lanosterol-labeling was about 7-fold higher inthe cholesterol-depleted compared to the cholesterol-rich cells.

From these data, it becomes clear that a block exists in thesterol synthesis pathway at the stage of lanosterol conversionto cholesterol. This also explains why neither mevalonic acid norsqualene nor lanosterol added to the culture medium had agrowth-stimulating effect. The data in Chart 4 further show that

the main compound of the isoprenoid pathway does not seemto be cholesterol but ubiquinone.

For a further characterization of the growth properties of these

OPTS

Charts. Combined growth curves (x x), thymidine incorporation rates(. .), and DNA histograms (right) of cells cultured under normal conditions (A),at 3% serum without cholesterol (B),and at 3% serum plus cholesterol (C).For theDNA measurements,between 8 and 23 x 103cells were analyzed.

JUNE 1984 2671

on March 5, 2020. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

E. W. Haeffner et al.

cells kept at a reduced serum concentration, we have measuredthe thymidine incorporation rate and also determined the amountof DNA within the cell cycle which is shown in Chart 5. Acomparison between the control cells (A) and the cells grown at3% serum plus cholesterol in the medium (C) reveals very similarpatterns not only in the cell density-dependent macromolecule

synthesis but also in the distribution of DNA between the d, S,and G2 phases. On the other hand, cells grown at 3% serumwithout cholesterol in the medium show a reduced growth rateresulting in the accumulation of cells in the mitotic and G2 phase.From these results, we can postulate that cholesterol must playa significant role in the expression of the growth properties ofthese ascites cells.

In the following, we have measured the cholesterol content,and the C/P ratio of the plasma membranes of control, cholesterol-enriched, and -depleted cells. As shown in Table 1, the cells

grown at 3% serum plus cholesterol contain about 30% morecholesterol, and the cells grown at 3% serum concentrationwithout cholesterol contain about 54% less cholesterol than thecontrol cells. Also, differences were observed in the amount ofphospholipids of these cell membranes, but their relative composition was mainly unchanged (Table 2). Determination of theC/P ratio as a measure of cell surface alterations gave values of0.29 for the control, 0.17 for the cholesterol-depleted, and 0.25for the cholesterol-rich cells. In order to correlate these data with

membrane fluidity changes within the lipid bilayer, we haveperformed fluorescence polarization measurements by using 1,6-diphenyl-1,3,5-hexatriene as membrane probe. The results of

these experiments (Chart 6) indicate some small differences inthe temperature-dependent fluorescence depolarization betweenthe cholesterol-rich and -poor cells in the higher temperature

range only.To further characterize the surface properties, we have tested

the capacity of these cells to agglutinate in the presence of

Table 1Cholesterol and phospholipid content, and the C/P molar ratio of the plasma

membranes of ascites cells grown under normal and reduced serumconcentrations (3%) without and with cholesterol supplementation

Data represent the average of at least 2 experiments with duplicate analyses.

Minimum essentialmedium+10%

serum+3% serum+3 % serum and cholesterol (2.50

mg/100ml)Cholesterol

(»g/mgprotein)19.6

9.025.2Phospho

lipid(ng/mgprotein)151.8

108.0256.0C/P(M)0.29

0.180.25

concanavalin A. The results of these experiments are shown inChart 7. Under the conditions applied, the agglutination occurredduring the first 3 to 8 min. However, the extent of agglutinationwas most dramatic with the cholesterol-enriched cells. Almost

90% of all cells were agglutinated after 2 min of reaction time,whereas the control showed only about 50%, and the cholesterol-poor cells, 30% agglutination. Since, in all these cases,

some agglutinates were present before the experiments werestarted, we have counted only the single cells from the beginningof the tests. Binding studies with labeled concanavalin A (Chart8) revealed some differences in the binding capacity of these

0.25 —i

o.zo

NIo:

0.20 —

0.15 —

OQ_

O

LJo</)LuorO

0.10 —

0.05 —

0.00

30 3.2

/ T (°K )3.410"

3.6

Chart 6. Temperature dependence of fluorescence polarization P of the plasmamembranes of cholesterol-rich (D d), cholesterol-poor cells (x x). Suspensions of 1.50 mg of membrane protein in 2.5 ml of Tris-buffered 0.25 M sucrose,pH 7.4, were labeled with 2 x 10"* M 1,6-diphenyl-1,3,5-hexatriene. Measurements

were performed with a Perkin-Elmer MPF 4 fluorospectrophotometer. Evaluationof the data obtained from 2 individual experiments was made by regression analysisyielding the equations y = 0.285 •10"3 x -0.8097 for the cholesterol-rich cells andy = 0.331 •10~3 x -0.9658 for the cholesterol-poor cells, with correlation

coefficients of r = 0.99 in both cases.

Table 2Phospholipid composition of horse serum and the plasma membranes of cholesterol-depleted and -enriched

ascites tumor cells

SerumUpidLysophosphatidylcholine

SphingomyelinPhosphatidylcholinePhosphatidylserinePhosphatidytethanolamineCardiolipinng

of phosphorus per 0.1 ml

of serum(31 ml ofmedium)%766.8

±480.8 ±

31 35.9 ±226.3 ±436.8 ±

0.0228a

208996

5519915.2

9.562.1

4.58.70Cholesterol-depleted

cellsng

phosphorus/mg of membrane

protein%885.9

±1798.9 ±

302.5 ±1251.3±

107.3±9.1

72.881.9

113.328.320.4

41.47.0

28.82.5Cholesterol-enrichedcellsng

phosphorus/mg of membrane

protein%2468.3

±3834.5 ±

811.5 ±2582.5 ±

0.0260.2

50.212.023.325.5

39.58.4

26.60

°Mean ±S.D.

2672 CANCER RESEARCH VOL. 44

on March 5, 2020. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Control of Growth and Surface Properties ofAscites Cells

4 6

MINUTES

Chart 7. Rate of agglutination of control cells (A A), of cells grown at 3%serumwithout cholesterol (• •),and of cells grown at 3% serum pluscholesterol(x x). Agglutination was induced with concanavalinA (1.0 mg per 10* cells in2 mlof phosphate-bufferedsaline,pH 7.4). Incubationwas performedundershakingat 37°.For evaluation, only single cells were counted at the beginning as well asduring the experiment. Curves, mean of 3 to 4 individualexperiments; oars, S.D.

40—i

800

Chart 8. Scatchard plots of binding of W-[acefy/-3H]concanavalinA to ascitescells grown in cholesterol-rich(x x) and cholesterol-poor medium(O O). B,amount of bound; F, amount of free ligand; B, nmols of concanavalinA bound perliter. Aliquots of 5 x 10s washed cells per ml were used. Points, average of 2experiments. Lines fit by regression analysis.

cells mainly at the low-affinity binding site. From the Scatchardplots, the high-affinity binding capacity was calculated to be 9.5x 10" for the cholesterol-depleted and 14 x 104 for the choles

terol-enriched cells with about the same association constant of2.4 x 107 x M~1.The low-affinity binding capacities were deter

mined with 1.3 and 2.5 x 106, respectively, with associationconstants of 4.2 and 1.1 x 106 x ivr1.

For further substantiation of the increased agglutinability ofthe cells at 3% serum plus cholesterol in the medium, we havecultivated the cells until they reached the plateau phase, andthen analyzed by interference microscopy under in vivo conditions. The results of these experiments are shown in Fig. 1. Incontrast to control cells and the cells kept without cholesterol,the cholesterol-enriched cells showed numerous aggregates in

various configurations. These aggregations seem to be formedby the cell plasma or microvilli building strong rod-like bridges

between cells.

DISCUSSION

It has been suggested that the primary role of serum is toprovide hormones for cell growth (23). Although the importanceof hormones with respect to the cultivation of cells seems to beundisputed, there have been found a number of low-molecular-

weight nutrients such as lipids and others which are equallynecessary for cell proliferation [see review by Barnes and Sato(6)]. Some of these substances, like phosphoethanolamine andselenic acid, have been tested by us in addition to the compoundsdescribed before, but without any effect. Our results clearlydemonstrate that ascites tumor cells need preformed cholesterolfor growth which normally is supplied by the serum. The observation that cells at 0.5% serum plus exogenous cholesterol inthe medium grow only to some extent indicates that othergrowth-promoting factors become limiting. These factors may

be different from those growth factors which we have added tothe medium and which have been shown to induce proliferationin ascites cells (30), in a human mammary tumor cell line (5), andin others (8). The block at the G2 phase of the cholesterol-

depleted cells has not been observed previously. However, ithas been known that the physical state of the lipid-protein matrix

of the plasma membrane undergoes profound changes afterstimulation of cell division and that the changes are correlated inone way or another to the presence of cholesterol (11). Theexact function of cholesterol in this respect is still open to debate(1,14). We are presently investigating the question in which waycell proliferation, surface function, and morphology may be regulated by the presence or absence of lipids.5

The regulation of cholesterol biosynthesis and the homeostaticcontrol of the cholesterol level are very complex mechanisms,still far from being resolved. For liver and also some hepatomas(3), it is known that the /3-hydroxymethylglutaryl coenzyme A

reducíase is the key enzyme in the sterol synthesis pathwaywhich is controlled through a multivalent feedback mechanism,involving cholesterol itself and possibly also a nonsterol compound (10). With respect to the involvement of cholesterol, thereare at least 3 independently operating negative-feedback mech

anisms, which have been described by Sabine (37). One of thesemechanisms, namely, the dietary feedback regulation, may be

5E. W. Haeffner,N. Pawletz, and C. J. K. Hoffmann, manuscript in preparation.

JUNE 1984 2673

on March 5, 2020. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

E. W. Haeffner et al.

functioning in our ascites cells, since acetate incorporation intocholesterol can be suppressed by a factor of about 7 at a highcholesterol level in the medium. Thus far, these ascites cellsdiffer from a number of extrahepatic tissues including rodenthepatomas (37, 41), which lack any control mechanism at all.Taking into account the low cholesterol-synthetic capacity of

these ascites cells, feedback control by endogenous cholesterolcan almost be excluded. It also remains to be investigatedwhether or not a nonsterol compound plays any role in thisrespect, or whether a second rate-limiting step exists besidesthe j3-hydroxymethylglutaryl coenzyme A reductase-controlled

reaction. The accumulation of lanosterol as the highest labeledproduct, which has also been observed for kidney cells (24),human blood leukocytes (12), and human lymphocytes andgranulocytes (13), seems to be indicative of a defect in thereaction sequence converting lanosterol to the end product. Thisdefective mechanism, however, would nicely explain the lowcholesterol de novo synthesis of these ascites cells.

The cultivation of our ascites cells under cholesterol-rich and-depleted conditions caused significant changes in the content

of this molecule as well as in the C/P ratio of the isolated plasmamembranes. The total lipid phosphorus content was also affected, but the relative proportions of the individual phospholipidswere not altered to a major extent between the 2 cell types. Theincrease in rigidity of liquid crystalline lipid phases induced bycholesterol has been shown by a number of physical measurements, and in the case of lymphocytes and lymphomas, thecholesterol content has been directly correlated with the degreeof fluorescence polarization (40). In spite of some changes in theC/P ratio between the plasma membranes of the cholesterol-richand cholesterol-poor cells, we could not find any significant

alterations in membrane lipid fluidity. There is some tendency fora rigidifying effect of the extra cholesterol in the higher temperature range beginning at about 30°.These results indicate that

these membranes apparently have the ability to control theirfluidity, e.g., by releasing phospholipids in case of a low cholesterol supplementation and wee versa. With respect to membranefluidity changes, it can further be generalized that these changesoccur primarily between C/P ratios of 0.2 and 0.5 (43), and thatthey are somehow correlated to the agglutinability of cells eitherin a cooperative (2, 4, 34, 36, 42) or in an antagonistic fashion(7, 28). Our study in which we used the most common probes,diphenylhexatriene and concanavalin A, to investigate the degreeof fluidity of the membrane lipid core as well as the degree ofmobility of membrane protein receptors, does not show a directcorrelation between these 2 phenomena. It remains to be established whether the concanavalin A-induced or spontaneous ag

glutination of these ascites cells is a direct effect of cholesterolitself, or a metabolite of cholesterol or of some other meansmediated by this molecule.

ACKNOWLEDGMENTS

The technical assistance of A. Roll in preparing the plasma membranefractionsand making the phospholipid analyses is greatly appreciated.

REFERENCES

1. Adam, G., Alpes, H., Blaser, K., and Neubert, B. Cholesteroland phospholipidcontent of 3T3 cells and transformed derivatives. Z. Naturforsch. Sect. C.Biosci., 30: 638-642, 1975.

2. Alderson, J. C. E., and Green, C. Lectin-induced cell agglutination and mem

brane cholesterol levels. Exp. Cell Res., 774: 475-479,1978.3. Avigan, J., Williams, C. D., and Blass, J. P. Regulation of sterol synthesis in

human skin fibroblast cultures. Biochim. Biophys. Acta, 2Õ8:381-384, 1970.4. Bales, B. L, Lesin, E. S., and Oppenheimer, S. B. On cell membrane lipid

fluidity and plant lectin agglutinability. A spin label study of mouse ascitestumor cells. Biochim. Biophys. Acta, 465: 400-407,1977.

5. Barnes, D., and Sato, G. H. Growth of a human mammary tumour cell line ina serum-freemedium. Nature (Lond.), 28Õ:388-389, 1979.

6. Barnes, D., and Sato, G. H. Methods for growth of cultured cells in serum-freemedium. Anal. Biochem., 702: 255-270, 1980.

7. Ben-Bassat, H., Polliak, A., Rosenbaum, S. M., Naparstek, E., Shouval, D.,and Inbar, M. Fluidity of membrane lipids and lateral mobility of concanavalinA receptors in the cell surface of normal lymphocytes and lymphocytes frompatients with malignant lymphomas and leukemias. Cancer Res., 37: 1307-1312, 1977.

8. Breitman, T. R., Collins, S. J., and Keene, B. R. Replacement of serum byinsulin and transferrin supports growth and differentiation of the human pro-myelocytic cell line, HL-60. Exp. Cell Res., 726: 494-498, 1980.

9. Brenneman, 0. E., McGee, L. and Spector, A. A. Cholesterol metabolism inthe Ehrlichascites tumor. Cancer Res., 34: 2605-2611,1974.

10. Brown, M. S., and Golstein,J. L. Multivalent feedback regulationof HMG CoAreductase, a control mechanism coordinating isoprenoid synthesis and cellgrowth. J. Lipid Res., 27. 505-517, 1980.

11. Burger, M. M. Cell surfaces in neoplastiatransformation. In: Current Topics inCellular Regulation, Vol. 3, pp. 135-193. New York and London: AcademicPress, Inc., 1971.

12. Bums, P., Welshman, I. R., Edmond J., and Spector, A. A. Evidence for rate-limiting steps in sterol synthesis beyond 3-hydroxy-3-methylglutaryl-coenzymeA reductase in human leucocytes. Biochim. Biophys. Acta, 572: 345-351,1979.

13. Bums, P., Welshman, I. R., Scallen, T. J., and Spector, A. A. Mechanism ofdefective sterol synthesis in human leucocytes. Biochim. Biophys. Acta, 773:519-528, 1982.

14. Chapman, D. Some recent studies of lipids, lipid-cholesterol and membranesystems. In: D. Chapman and D. F. H. Wallach (eds.). Biological Membranes,Vol. 2, pp. 91-144. London: Academic Press, Inc., 1973.

15. Corwin, L. M., Humphrey, I. P., and Shloss, J. Effect of lipidson the expressionof cell transformation. Exp. Cell Res., 708: 341-347,1977.

16. EiW,H. J., and Lans, W. E. M. A new, sensitive determination of phosphate.Anal. Biochem., 30: 51-57, 1969.

17. Feller, M., Richardson, C., Behnke, W. D., and Gruenstein, E. High and lowaffinity binding sites for concanavalin A on normal human fibroblasts in vitro.Biochem. Biophys. Res. Commun., 76: 1027-1035, 1977.

18. Folch,J., Lees, M., and Sloane-Stanley,G. H. A simplemethod for the isolationand purification of total lipids from animal tissues. J. Biol. Chem., 226: 497-509,1957.

19. Fuchs, P., Parola,A., Robbins, P. W., and Btout, E. R. Fluorescencepolarization and viscosities of membrane lipids of 3T3 cells. Proc. Nati. Acad. Sci.USA, 72:3351-3354,1975.

20. Haeffner,E. W., Kolbe, K., Schroeter, D., and Paweletz, N. Plasmamembraneheterogeneityin ascites tumor cells. Isolationof a light and a heavy membranefraction of the glycogen-free Ehriich-Lettrésubstrain. Biochim. Biophys. Acta,603:36-51,1980.

21. Haeffner, E. W., and Hoffmann, C. J. K. Direct quantitation of free cholesterolfrom total serum lipid extracts by computer-assisted gas liquid chromatogra-phy. J. Chromatogr., 228: 268-272, 1982.

22. Hate,A. H., Pessin,J. E., Palmer,F., Weber, M. J., and Glaser,M. Modificationof the lipid composition of normal and Rous sarcoma virus-infected cells. J.Biol. Chem., 252. 6190-6200, 1977.

23. Hayashi,I., and Sato, G. H. Replacementof serum by hormonespermit growthof cells in a defined medium. Nature (Lond.). 259: 132-134, 1976.

24. Hellstrom, K. H., Sipperstein, M. D., Bricker, L. A., and Luby, L. J. Studies ofthe in vivo metabolism of mevalonicacid in the normal rat. J. Clin. Invest., 52:1303-1313,1973.

25. Hoffmann, C. J. K., and Haeffner, E. W. Effect of low serum concentrationwith and without lipid supplementationon Karzel ascites tumor cell growth. J.Cancer Res. Clin. Oncol., 99: A11-A12,1981.

26. Hoffmann, C. J. K., Paweletz, N., Friedel, R., and Haeffner, E. W. Structuraland functional alterations of lipid-depleted ascites tumor cells in culture. Eur.J. Cell Biol., 33: 66-74,1984.

27. Hostek, H. L. Uptake and utilization of free fatty adds supplied by liposomesto mammarytumor cells in culture. Exp. Cell Res., 722:127-136,1979.

28. Inbar, M., Shinitzky, M., and Sachs, L. Rotational relaxation time of concanavalin A bound to the surface membrane of normal and malignant transformedcells. J. Md. Biol., 87: 245-253,1973.

29. Kahane, I., and Razin, S. Cholesterol-phosphatidylcholmedispersions as donors of cholesterol to Mycoplasma membranes. Biochem. Biophys. Acta, 477:32-38, 1977.

30. Lehmann,W., Graetz, H., Schutt, M„and Langen, P. Antagonistic effects ofinsulin and a negative growth regulator from ascites fluid on the growth ofEhrlichascites carcinoma cells in vitro. Exp. Cell Res., 779: 396-399, 1979.

31. Mamaril, F. P., Dobrjansky, A., and Green, S. A rapid method for the isolationof nuclei from Ehrlich ascites tumor cells. Cancer Res., 30: 352-356,1970.

2674 CANCER RESEARCH VOL. 44

on March 5, 2020. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Control of Growth and Surface Properties of Ascites Cells

32. Monard, D., Rentsch, M., Schuerch-Ralhgeb, Y., and Lindsay, R. M. Morpho- Acad. Sci., 57: 660-672,1949.logical differentiation of neuroblastoma cells in medium supplemented with 39. Shinitzky, M., and Barenholz, Y. Dynamics of the hydrocarbon layer in lipc-delipidated serum. Proc. Nati. Acad. Sci. USA, 74:3893-3897,1977. somes of lecithin and sphingomyelin containing dicetylphosphate. J. Biol.

33. Odriozda, J. M., Waitzkin, E., Smith, T. L, and Bloch, K. Sterol requirement Chem., 249: 2652-2657,1974.of Mycoplasma capricolum. Proc. Nati. Acad. Sci. USA,75:4107-4109,1978. 40. Shinitzky, M., and Inbar, M. Difference in microviscosity induced by different

34. Rittenhouse, H. G., Williams, R. E., Wisnieski, B., and Fox, C. F. Alterations cholesterol levels in the surface membrane lipid layer of normal lymphocytesof characteristic temperature for lectin interactions in LM cells with altered lipid and malignant lymphoma cells. J. Mol. Biol., 85: 603-615,1974.composition. Biochem. Biophys. Res. Commun., 58: 222-228,1974. 41. Sipperstein, M. D., Gyde, A. M., and Morris, H. P. Loss of feedback control of

35. Röschlau,P., Bemt, E., and Gruber, W. Enzymatische Bestimmung des hydroxymethylglutarylcoenzyme A reducíasein hepatomas. Proc. Nati. Acad.Gesamt-Cholesterinsim Serum. Z. Klin. Chem. Klin. Biochem., 12: 403-407, Sci. USA,68: 315-317,1971.1974. 42. Tombaccini, D., Ruggieri, S., Fallani, A., and Mugnai, G. Concanavalin A-

36. Rule, G. S., Kruuv,J., and Lepock, J. R. Membranelipid fluidity as rate limiting mediated agglutinability of BALB/c 3T3 cells grown in media supplementedin the concanavalinA-medialedagglutinationof pyBHK cells. Biochim.Biophys. wilh differenl phosphatidylcholines. Biochem. Biophys. Res. Commun., 96:Acta, 556: 399-407, 1979. 1109-1115,1980.

37. Sabine, J. R. Defective control of cholesterol synthesis and the development 43. Wallach, D. F. H. Cholesterol. In: D. F. H. Wallach (ed.), Membrane Molecularof liver cancer: a review. In: R. Wood (ed.), Tumor Upids: Biochemistry and Biology of Neoplastic Cells, pp. 217-241. Amsterdam: Elsevier ScientificMetabolism,pp. 21-33. Champaign, IL: AmericanOil Chemisls' Society, 1973. PublishingCo., 1975.

38. Scalchard, G. The attraction of proteins for small moleculesand ions. Ann. NY

JUNE 1984 2675

on March 5, 2020. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

E.W. Haeffner et al.

Fig. 1. Interference phase contrast microscopy of cholesterol-rich and -poor cells. Cells were kept in minimum essential medium to which 3% horse serum andcholesterol (2.50 mg/100 ml) were added. Cell cultures were kept for 4 days without diluting the cell number but with one medium change. At the time of the analysis,the cell density was 8.6 x 105 cells/ml. The cells were carefully placed on glass plates and observed immediately under the microscope; a and b, cholesterol-depletedcells; c and d, cholesterol-enriched cells; magnification: a and c, bar equals 50 i*m; b and d, bar equals 12.5 pm.

2676 CANCER RESEARCH VOL. 44

on March 5, 2020. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1984;44:2668-2676. Cancer Res   E. W. Haeffner, C. J. K. Hoffmann, M. Stoehr, et al.   Membrane Properties of Ascites Tumor Cells in CultureCholesterol-induced Growth Stimulation, Cell Aggregation, and

  Updated version

  http://cancerres.aacrjournals.org/content/44/6/2668

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/44/6/2668To request permission to re-use all or part of this article, use this link

on March 5, 2020. © 1984 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from