Differences in Lipid Characteristics of Undifferentiated and Enterocytic-differentiated HT29 Human Colonic Cells

Max Reynier, Hlne Sari, Manola d'Anglebermes, et al. Cancer Res 1991;51:1270-1277.

Updated version

Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/51/4/1270

E-mail alerts Reprints and Subscriptions Permissions

Sign up to receive free email-alerts related to this article or journal. To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at pubs@aacr.org. To request permission to re-use all or part of this article, contact the AACR Publications Department at permissions@aacr.org.

Downloaded from cancerres.aacrjournals.org on October 12, 2013. 1991 American Association for Cancer Research.

(CANCER RESEARCH 51. 1270-1277, February 15. 1991]

Differences in Lipid Characteristics of Undifferentiated differentiated HT29 Human Colonie Cells1

and Enterocytic-

Max Reynier,2 Hlne Sari, Manola d'Anglebermes, Edouard Ah Kye, and Lyda Pasero
Institut de Chimie Biologique, CNRS-U&4 202, Universit d'Aix-Marseille I, 3 place Victor Hugo, I3331 Marseille cedex 3, France

ABSTRACT
In this paper we compared several lipid characteristics of the homogenate and the corresponding plasma membrane in Undifferentiated and differentiated HT29 human colon cancer cells, using normal human colonie cells as a reference. Electron microscopy showed that HT29 cells were morphologically Undifferentiated when cultured in the presence of either glucose or inosine without glucose at early confluency. On the contrary, HT29 cells cultured at late confluency in a glucose-free medium containing inosine or grown in nude mice exhibited an enterocytic differ entiation with the presence of tight junctions and an apical brush border. The cell homogenate and the plasma membrane were prepared from each cell type. The study of specific marker enzymes showed the same degree of purity in all plasma membranes, with a highly marked increase of brush border-associated hydrolases (A'-aminopeptidase and alkaline phosphatase) only in the organelles isolated from differentiated HT29 and colonie cells. Respective similar increases in the amount of free cholesterol and phospholipid and in the free cholesterol:phospholipid molar ratio were found in the plasma membrane as compared with the homogenate in all HT29 cell types. This ratio, due to an increased phospholipid content in both homogenate and plasma membrane, was lowered in colonie cells. No differences in the phospholipid profile were found between the homogenates of all cell types and the plasma mem brane of Undifferentiated HT29 cells, with the exception of a decrease of cardiolipin in this organelle. On the contrary, the plasma membrane phospholipid composition was different from that of the corresponding homogenate in differentiated HT29 and colonie cells. The most striking changes were a highly increased sphingomyelin amount and concomitant decreases in phosphatidylethanolamine, phosphatidylserine, and cardi olipin. Moreover, differences in the percentage of phosphatidylcholine plus sphingomyelin as well as in phosphatidylcholine:sphingomyelin, phosphatidylethanolamine, and/or phosphatidylcholine molar ratios were also found. The monounsaturated:polyunsaturated fatty acid ratio in phosphatidylethanolamine was similar in differentiated HT29 and colonie cells and lower than in Undifferentiated HT29 cells. A decrease in this latter ratio in phosphatidylcholine was also observed in colonie cells and HT29 cells grown in nude mice. These changes were essentially due to opposite variations in the percentage of palmitoleic acid and those of linoleic and/or arachidonic acids in both phospholipids. Thus, these data indicate that Undifferentiated HT29 cells were characterized by the absence of a specific phospholipid composition in their plasma membrane, which is suggested to be related to altered phospholipid sorting. The plasma membrane phospholipid profile reversed essentially to the normal pattern when HT29 cells recovered the ability to differentiate.

is still uncertain, and no general trend is discernible concerning phospholipid modifications. Whether restoration of ultrastruc tural characteristics during the differentiation of tumor cells is accompanied by the regression of biological and biochemical features, typical of a malignant phenotype, is a question still unanswered (6, 12, 13). The human colon adenocarcinoma cell line HT29 appears to be a valuable experimental model to determine the respective influence of the malignant transformation and differentiation in the expression of cellular constituents and functions of the intestinal epithelium, since these cells are available under either an Undifferentiated or a differentiated state (14-18). The differ entiation process of this cell line is reversible and mimics the normal intestinal ontogenesis and epithelial crypt-to-villus mi gration. However, these cells are malignant and of colonie origin, and the differentiated HT29 cells exhibit the closest analogies to human fetal colon (16, 19, 20). In spite of these limitations, recent reports are indicative of the potential interest in this experimental model (21-27) to determine the respective influence of neoplasia and differentiation in the expression of morphological and structural events. Lipids are involved in a broad spectrum of cellular functions (10, 11). In particular, changes in the total cholesterol: phospholipid molar ratio as well as in the relative percentage and degree of fatty acid saturation of phospholipids have been found to influence a number of important functions in the rat intestinal antipodal plasma membranes (28-30). In this study, we compare the phospholipid composition of the whole cell homogenate and the corresponding plasma mem brane fraction of Undifferentiated and differentiated HT29 cells with human normal mature colonie cells used as a reference. The most striking change was that the specific phospholipid profile in the plasma membrane of normal colonie cells did not occur in Undifferentiated HT29 cells but was recovered in differentiated cells, suggesting that the phospholipid polarized delivery to the cell surface is severely altered in relation to the neoplastic process. MATERIALS AND METHODS
Cultured Cells, Tumors in Nude Mice, and Normal Adult Colonie Cells. The established cell line HT29 (HT29-Glc), which was derived from a human colon adenocarcinoma (31), was cultured in Dulbecco's modified Eagle's medium (Gibco-Brl, Cergy Pontoise, France) contain ing 25 HIMglucose and supplemented with 10% (v/v) PCS' (SeralabBiosys, Compigne, France) (15, 18). These cells were adapted in one step to grow in the same medium devoid of glucose and pyruvate (Eurobio, Paris, France), supplemented with 2.5 mM inosine (Sigma Chimie, La Verpillire, France) and 10% dialyzed FCS. Under these conditions, no significant cell loss was observed and the resulting HT29 cells (HT29-Ino) exhibited a confluency-dependent enterocytic differ entiation as previously reported (17, 18, 22). Both cell types were routinely subcultured by trypsinization (0.25% trypsin plus 0.53 mM
3The abbreviations used are: PCS, fetal calf serum; DMA, dimethylacetals; PBS, phosphate-buffered saline: TLC. thin-layer chromatography.

INTRODUCTION Various attempts have been performed to determine the possible changes in phospholipid patterns associated with ma lignancy (1-5) and/or differentiation (6-9) and to correlate these variations with the organization and functioning of cell ular membranes (10, 11). However, the relevance of these data
Received 6/14/90: accepted 11/27/90. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1This work was supported in part by grants from FNCLCC and CNRS and was performed in CNRS-URA 202 as directed by Dr. J. Marvaldi. ! To whom requests for reprints should be addressed.

1270

LIPID CHANGES DURING DIFFERENTIATION

OF TUMOR CELLS

EDTA in PBS) every week. The cells were seeded at 10 x 10" cells/ 150-cm2 plastic flask and maintained at 37Cin a humidified atmos phere of 95% air and 5% CO2. The culture medium was renewed daily. The experiments were performed at days 24 to 28 after seeding for HT29-G1C cells. The HT29-Ino cells (passages 7 to 12) were used at early (18 to 20 days in culture) or late (30 to 32 days in culture) confluency and will be referred to as early confluent HT29-Ino cells and late confluent HT29-Ino cells, respectively. Cell layers were rinsed 3 times with PBS, pH 7.2, and immediately fixed in situ for electron microscopy or mechanically scraped for biochemical analyses. HT29N cells were obtained as previously reported (14, 16, 32) from tumors developed in athymic nu/nu mice (Cseal-Cnrs, Orlans, France) after s.c. injection of 10 x 10' HT29-Glc cells into animals fed under conventional sterile conditions. After 30 to 40 days of growth, solid tumors were excised immediately after decapitation, the connective tissue was discarded, and the tumors were minced with surgical scissors. Normal adult colonie cells were scraped, using a glass microscope coverslip from fresh surgical colon fragments extensively prewashed with 0.24 M NaCl. The dislodged cells were epithelial cells as found by cytological studies and as reported previously (33). Transmission Electron Microscopy. Cell layers in the culture flasks or tumor samples were fixed for l h with 2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4, and then with 1% osmium tetroxide in 0.2 M phosphate buffer. After extensive washing with the previous buffer, the fixed samples were dehydrated in graded alcohols and embedded in Epon. Thin sections were cut with a LKB 2088 ultratome, contrasted with uranyl acetate and lead citrate, and examined with a Jeol 100 C electron microscope. Preparation of Whole Cell Homogenate and Plasma Membrane Frac tion. Plasma membrane fractions were isolated essentially according to Warren (34) and Perdue (35). All procedures were carried out at 4C. Cells were washed 3 times with 0.16 M NaCl and twice with 0.05 M Tris-HCl, pH 7.4, by resuspension and centrifugation. The cell pellets were suspended at a concentration of 1 x IO7 cells/ml in 0.005 M MgCl2:0.05 M Tris-HCl, pH 7.4, allowed to stand for 10 min, and disrupted by 20 gentle strokes in a Dounce homogeneizer. This disrup tion step was repeated once more for human normal colonie cells. The whole-cell homogenates were sedimented at 105,000 x g for 1 h, and the resulting pellets were homogeneized in 20% sucrose (w/w). Four ml of this particulate suspension were layered onto a 30-ml continuous density gradient of sucrose (20 to 45%, w/w) with 2 ml of 50% sucrose solution on the bottom. After 15 h of centrifugation at 25,000 rpm in a SW27 Beckman rotor, plasma membranes were found at a band of density 1.08 to 1.15 on the basis of enzymatic criteria. This layer was harvested, diluted with 0.16 M NaCl, and sedimented at 105,000 x g for 45 min. The amount of protein recovered in this plasma membrane fraction represented about 2.5% of the cell homogenate protein, which is usual for plasma membrane preparation from various cell lines (36, 37). Enzymatic Assays. Ouabain-sensitive, (Na+,K+)-dependent ATPase

by 2-h stirring in 20 volumes of chloroform:melhanol (2:1, v/v) al room temperature according to the technique of Folch et al. (45). After partition with 0.2 volume of 0.85% KCI, the lower organic phase was evaporated to dryness, resuspended in a small volume of chloroform:melhanol (2:1, v/v), and stored al 20C under nitrogen. Free cholesterol was estimated using an enzymatic melhod utilizing choles terol oxidase, catalase, and acetylacelone (Boehringer-Mannheim, Meylan, France). Phosphorus in phospholipid was eslimated by the method of Ames (39). Phospholipids (150 nmol) were fractionated by two-dimensional TLC on 20-x 20-cm precoated silica gel H plates (Merck, Nogent-surMarne, France) using chloroform:acetone:methanol:acelic acid:waler (50:20:10:15:5,v/v) in ihe firsl dimension and chloroform:methanohpetrol etheracetic acid:boric acid (40:20:30:10:1.8, v/v/v/v/ w) in ihe second dimension. Phospholipids were delected by spraying with a specific reagent (46) and idenlified by comparison wilh reference slandards. After Ihe first migration, plates were sprayed with 0.005 M HgCU in 0.1 M acetic acid and dried under vacuum for l h for plasmalogen determinalion. Individual spols were scraped off and analyzed for phosphorus coment as described above. Fally acid analyses of phosphalidylcholine and phosphalidylelhanolamine isolated from preparative TLC of cell homogenate phospholipids (750 nmol) were performed as described earlier (5). Fatly acid melhyl eslers and DMA of both isolated phospholipids were prepared by boron Irifluoride Iransmelhylalion and analyzed isolhermally al 185C by gas-liquid chromalography wilh a Varian 1440 gas chromalograph equipped with a hydrogen flame ioni/ai ion detector in a column of 7% butanediolsuccinale on Gas-Chrom Q, 100/200 meshes (Gerdel, Suresnes, France) wilh a nitrogen flow rate of 28 ml/min. Peaks were identified by comparison wilh ihe relenlion lime of reference slandards and were quaniiiated by integralion of iheir areas wilh a Varian CDS III calculator.

RESULTS

Ultrastructural Characterization of HT29 Cells Grown in Cul ture and in Nude Mice Tumors. HT29-Glc cells at late confluency displayed disorganized multilayers of undifferentiated cells (Fig. IA). These cells exhibited cytoplasmic projections randomly dispersed at the cell periphery and interacted by numerous desmosomes but lacking tight junctions. Early con fluent HT29-Ino cells formed essentially a monolayer with some multilayered regions of unpolarized cells (Fig. B). Only spare, irregular microvilli were present on the surface layer with few tight junctions. At late confluency, HT29-Ino cells were organized into a typical polarized monolayer with well-devel oped junctional complexes and characteristic apical brush bor der microvilli (Fig. 1C). HT29-N cells showed the presence of (EC 3.6.1.3) activity was measured according to the method of Gratecos intercellular cysts with a regular and well-organized brush bor et al. (38). 5'-Nucleotidase (EC 3.1.3.5) activity was determined by der in the majority of them (Fig. \D). liberation of inorganic phosphate (39). A'-Aminopeptidase (EC 3.4.11.2) Enzymatic Characterization of the Homogenates and Plasma and alkaline phosphatase (EC 3.1.3.1) activities were measured spectrophotometrically using appropriate substrates (40, 41). Rolenone-insenMembrane Fractions from Differentiated and Undifferentiated sitive NADPH-cytochrome c reduc-ase(EC 1.6.99.3) was detected by Cells. The specific activity of several marker enzymes was determined in both the whole-cell homogenate and the plasma the reduction of oxidized cytochrome c at 550 nm (42). Measurement ofcytochromecoxidase(EC 1.9.3.1) was done according to the method membrane fraction prepared from each cell type. The results of Cooperstein and Lazarow (43). Protein content was determined by are presented in Table 1. For each plasma membrane marker, the method of Lowry et al. (44) using bovine serum albumin as a the respective specific activities were nearly similar in all ho standard. The enzyme activities are expressed as milliunits per mg of mogenates. These values were markedly increased in the cor protein. One unit is defined as the activity that utilizes 1 ^mol of responding plasma membrane fractions except for both brush substrate per min under experimental conditions. Alkaline phosphatase, border-associated hydrolases in the morphologically undiffer 7V-aminopeplidase, and 5'-nucleotidase were utilized as apical mem entiated cells. In HT29-G1C and early confluent HT29-Ino cells, brane markers; (Na+,K+)-ATPase for the basolateral membrane; and the activities of the A'-aminopeptidase and alkaline phosphatase cytochrome c oxidase and NADPH-cytochrome c reduc-asefor the were, respectively, either only moderately or not significantly mitochondria and the endoplasmic reticulum, respectively. enhanced in plasma membranes compared with homogenates. Extraclion and Analyses of Lipids. Lipids from cell homogenates and plasma membrane fractions were extracted as previously described (5) On the other hand, plasma membrane fractions were depleted 1271

LIPID CHANGES DURING DIFFERENTIATION

OF TUMOR CELLS

-bb

Fig. 1. Ultrastructural characterization of HT29 cells grown in vitro and in nude mice. A, HT29 cells grown in the presence of 25 HIM glucose (28 days after seeding): vertical section showing a multilayer of undifferentiated cells with desmosomes (d) and cytoplasmic processes (cp) (X 3,600). B, early confluent HT29 cells grown in the absence of glucose and in the presence of 2.5 mM inosine (19 days after seeding): vertical section showing a monolayer of predifferentiated cells with irregular and sparse microvilli (m) and tight junctions ((/') (x 6,000). C, HT29 tumors in nude mice (36 days after injection): cross-section showing the presence of a regular brush border (bb) lining the luminal membrane of cells organized around intercellular crypts (x 20,000). D, late confluent HT29 cells grown in the presence of 2.5 TOM inosine and absence of glucose (30 days after seeding): vertical section showing a monolayer of polarized and differentiated cells with apical brush border microvilli (bb), desmosomes (</), and tight junctions (tj) (x 5,000). Inset, cross-section of brush border microvilli. Note, as in C, the two leaflets of the plasma membrane with the outer leaflet coated with filamentous material and the core of microfilaments which extend into the cytoplasm (x 20,000).

Table 1 Distribution of marker enzymes in whole-cell homogenates and plasma membrane fractions Preparation of whole-cell homogenates and plasma membrane fractions as well as the enzymatic assays was described in "Materials and Methods." Cells HT29-GIC confluentMarker enzymes(Na\K*)-ATPase Early HT29-Ino confluentWhole-cell HT29-N Normal colonie

5'-Nucleotidase jV-Aminopeptidase Alkaline phosphatase Cytochrome c oxidase NADPH cytochrome c reduc-aseWhole-cell Mean SD of 3 to 5 separate independent experiments, milliunits/mg of protein. One unit is defined as the activity that utilizes 1 jmol of substrate/min under the experimental conditions. * Values obtained in duplicate from one experiment. c ND, not determined.

homogenate36.4 membrane189.2 homogenate55.2* membrane152.3* homogenate46.0 membrane134.7 homogenate50.9 membrane140.2 homogenate43.5 membrane181.4 + 9.9 33.1 12.1 22.1 40.5 12.4 28.9 NDC 53.8 13.0 247.8 45.1 44.7 14.8 181.2 + 22.4 ND 32.6 0.8 158.9 6.6 30.2 6.6 128.9 19.2 15.3 + 2.4 67.2 + 25.1 5.3 1.9 18.2 2.6 6.1 1.9 6.0 1.4 39.6 2.6 7.1 1.0 108.9 37.3 11.9 3.5 1.0 + 0.3 7.0 + 2.1 1.3 + 0.2 6.4+ 1.2 0.6 0.2 1.8 1.1 2.1 1.3 1.3 0.2 5.1 2.4 1.6 0.4 11.8* 2.9* 10.3 + 2.1 4.9 + 2.1 15.6 + 2.4 6.2 1.2 2.0 1.3 2.1 1.5 16.2 3.9 0.5 0.3 3.2Plasma 1.2Late 4.9 + 2.7Plasma 1.5 + 0.7Whole-cell 1.8+ 1.1Plasma 4.814.6 1.4Plasma 0.7 0.2Whole-cell 2.4 0.3Whole-cell 5.2 0.9Plasma 2.3 1.1

of cytochrome c oxidase and NADPH-cytochrome c reduc-ase, and the contamination by mitochondria and endoplasmic reticulum was calculated to be less than 5% (data not shown).

Comparative Free Cholesterol and Phospholipid Contents of Cell Homogenates and Plasma Membrane Fractions. As shown in Fig. 2, the free cholesterohphospholipid ratio was higher in

1272

LIPID CHANGES DURING DIFFERENTIATION

OF TUMOR CELLS

] Plasma membrani Cell homogenate

a b c d e Free cholesterol

b c d Phospholipid

a b c d e Free cholesterol/phospriolipid

Fig. 2. Free cholesterol and phospholipid contents of whole-cell homogenates and plasma membrane fractions. Free cholesterol and total lipid phosphorus were estimated as described in "Materials and Methods." From these values, the free cholesterol:phospholipid ratios were calculated, a, HT29-Glc; >, early confluent HT29-Ino; c, late confluent HT29-Ino; d, HT29-N; e, normal colonie cells. Columns, mean; bars, SD (n = 3 to 5). *, Significantly different from respective values found in e with P < 0.05 (determined by Cochran's / test).

the plasma membrane fraction than in the cell homogenate for each cell type. Moreover, this ratio in both cell homogenates and plasma membrane fractions was, respectively, similar in all undifferentiated and enterocytic-differentiated HT29 cells and was enhanced as compared with that of normal colonie cells. This increase was due to the reduced amount of phospholipid in all types of HT29 cells, whereas the free cholesterol content was similar in these and normal colonie cells. Phospholipid Composition of the Cell Homogenates and Plasma Membrane Fractions. Except for cardiolipin which was lowered in all plasma membrane fractions, comparison of the phospholipid composition between the cell homogenate and the corresponding plasma membrane fraction was allowed to dis tinguish two groups among the studied cells according to the differentiation state (Table 2). In both HT29-Glc and early confluent HT29-Ino cells, the percentage distribution of phos phorus among the major phospholipid classes was similar in the cell homogenate and plasma membrane fraction. On the contrary, the plasma membrane phospholipid composition was different from that of the cell homogenate in late confluent HT29-Ino, HT29-N, and normal colonie cells. The most strik ing changes were the approximately 3-fold increase in the amount of sphingomyelin and a concomitant slight decrease of phosphatidylethanolamine in the plasma membrane fraction in

comparison to the respective cell homogenate. Moreover, there was a 2-fold decrease in the proportion of phosphatidylserine in the plasma membrane fraction of HT29-N and normal colonie cells, whereas such a change was not found in late confluent HT29-Ino cells. Table 2 also shows that the percent age of the two choline-containing phospholipids, which repre sented about 50% of the total phospholipid in all cell homoge nates, was unchanged in the plasma membrane fractions of both undifferentiated cells but was clearly enhanced in those of differentiated cells. Similarly, the phosphatidylcholine:sphingomyelin and phosphatidylcholine:phosphatidylethanolamine ra tios in the plasma membrane fractions were, respectively, greatly decreased and moderately increased in the three differ entiated cell types, while the phosphatidylcholine:phosphatidylserine ratio was only enhanced in plasma membrane frac tions of HT29-N and normal colonie cells. Unsaturated Fatty Acid Composition of Phosphatidylcholine and Phosphatidylethanolamine. The unsaturated fatty acid pat terns of phosphatidylethanolamine and phosphatidylcholine purified from cell homogenates are reported in Tables 3 and 4, respectively, and the ratio of monounsaturated to polyunsaturated fatty acids in both phospholipids is shown in Fig. 3. When compared with normal colonie cells, both undifferentiated HT29 cells were similarly characterized by a high proportion of palmitoleic acid in both phospholipids associated with low percentages of either linoleic, linolenic, and arachidonic acids in phosphatidylethanolamine or linoleic and arachidonic acids in phosphatidylcholine. Consequently, the monounsaturated: polyunsaturated fatty acid ratios in both undifferentiated HT29 cells were similarly increased 2- to 4-fold in phosphatidyletha nolamine and 6- to 8-fold in phosphatidylcholine as compared with those found in normal colonie cells. In both differentiated HT29 cells, the monounsaturatedrpolyunsaturated fatty acid ratio of phosphatidylethanolamine was similarly decreased as compared with that found in undifferentiated cell types and had the same value as in normal colonie cells. This reversion to the normal was due to a partial decrease in the palmitoleic acid proportion coupled with an increase only in the relative per centage of either arachidonic acid in late confluent HT29-Ino cells or both linoleic and linolenic acids in HT29-N cells. The unsaturated fatty acid pattern of phosphatidylcholine exhibited similar changes in HT29-N cells but was unmodified in late confluent HT29-Ino cells as compared with undifferentiated

Table 2 Phospholipid composition of cell homogenates and plasma membrane fractions Phospholipids in the organic phase of Folch's extracts from cell homogenates and plasma membrane fractions were analyzed by two-dimensional TLC and were quantitated by estimation of the inorganic phosphate. Cells HT29-G1C Early confluent HT29-lno Late confluent HT29-N Normal colonie

Whole-cell Whole-cell Whole-cell Whole-cell Whole-cell Plasma Plasma Plasma Plasma Plasma Phospholipids membrane membrane homogenate membrane homogenate homogenate homogenate membrane homogenate membrane 2.8*5.2 PC"SMPEPIPSCLPC+SMPC/SMPC/PEPC/PS48.5 1.16.4 1.03.3 0.45.4 0.83.2 1.5 2.3 5 1.9 0.9 43.2 0.9 44.3 0.4 46.3 0.3 5 1.3 0.4'27.9 0.7 15.7+1.0' 3.9+1.8 14.6 0.6' 5.3 + 0.6 14.8 + 1.622.8 0.622.3 0.827.4 1.226.8 19.7+1.9'6.2+1.4 1.6 20.1 + 0.9' 34.8 2.6 25.31.3C 27.2 2.4 4.36.7 0.58.9 1.28.3 3.99.6 0.47.2 7.6 0.8 8.0 + 0.8 7.9 + 0.7 7.6 + 0.8 8.2 1.29.5 1.012.3 0.76.6 1.07.8 0.4'3.8 + 0.6 5.5+1.3 6.0 + 0.7 3.2 1.2' 6.3+1.1 2.9 1.55.7 2.11.6 0.74.2 1.11.8 0.7'53.1 + 0.8'52.4 0.0'54.7 + 0.9 0.0' 2.8+1.0 0.8 0.3' 6.7 0.5 2.053.7 1.252.8 1.8'16.1 2.4 67.6+1.2' 47.1 + 2.8 58.9+1.1' 51.6+1.1 66.1 1.28.7 3.59.3 1.27.3 2.315.0 1.4'1.7 0.2'1.8 0.4 3.3 0.2' 11.1 0.5 3.0 + 0.3' 8.7+1.1 3.5 1.72.1 0.72.1 0.61.8 0.3'7.1 0.2 2.6 0.1' 1.2 0.1 1.8 + 0.1' 1.7 0.2 2.6 0.15.1 0.13.8 0.27.5 0.06.0 0.7 9.4+1.6 7.2 1.0 13.8+1.4' 7.3 1.3 17.7+1.4' 1.146.7 0.749.5 0.947.0 0.95 " PC, phosphatidylcholine; SM, sphingomyelin; PE, phosphatidylethanolamine + plasmalogens; PI, phosphatidylinositol: PS, phosphatidylserine; CL, cardiolipin. * Mean SD (n = 3 to 5), expressed as molar percentage of total phospholipid. From these values, the mean SD of the sum of the percentages of PC + SM and the ratios of the percentage of PC to that of SM, PE, or PS were calculated. ' P < 0.05 for values between plasma membrane fraction and cell homogenate of the same cell type (determined by Cochran's t test).

1273

LIPID CHANGES DURING DIFFERENTIATION

OF TUMOR CELLS

Table 3 Unsaturated fatly acid and DMA composition of phosphatidylethanolamine Fatty acid methylesters and DMA of phosphatidylethanolamine purified from cell homogenates were prepared by boron trifluoride transmethylation and ana lyzed by gas-liquid chromatography as described in "Materials and Methods." Cells HT29-Ino

components was unchanged after 3 days of cell growth without medium change (data not shown). DISCUSSION

This investigation revealed marked differences in the phospholipid pattern between undifferentiated and differentiated Fatty acids16:1 confluent21.4 confluent6.2 colonie3.1 HT29 human colonie adenocarcinoma cells. The phospholipid 2.6"39.3 2.0*26.6 0.7*24.118.91.61.28.42.09.61.81.9*0.3*0.20.70.41.5Normal 0.5*29.4 composition of the plasma membrane in undifferentiated cells u)718:1 0.732.21.21.26.42.37.62.60.30.20.81.21.3Late o)918:2^)618:3 4.33.3 3.51.4 2.612.4 resembled that of the whole cell homogenate. On the contrary, 2.8*2.9 1.20.3 0.30.7 0.5*2.2 significant differences in the phospholipid composition between 0.32.7 o)320:3 0.12.4 u)620:4 0.37.9 1.912.9 0.314.9 the cell homogenate and the plasma membrane found in normal 0.3*3.3 3.0*1.6 0.620:5 1.83.8 mature colonie cells occurred in differentiated HT29 cells. w3DMAHT29-G1C17.1 0.16.0 1.68.0 0.115.6 1.1* Moreover, changes in the monounsaturated:polyunsaturated 2.7Early 0.9HT29-N8.2 Mean SD of 3 to 5 determinations, expressed as weight percentage of total fatty acid ratio in phospholipids were also found to be related to the differentiation state of this cell line. fatty acids. * P < 0.05 compared with values in HT29-Glc cells (determined by Cochran's As previously reported (14-18), the present ultrastructural t test). data showed that HT29 cells cultured either in the presence of glucose (15, 18) or at early confluency in a glucose-free medium containing inosine (17, 18) were essentially undifferentiated. Table 4 Unsaturated fatty acid composition ofphosphatidylcholine Phosphatidylcholine purified after TLC of cell homogenate phospholipids was On the contrary, HT29 cells grown in nude mice (14, 16, 32) treated by boron trifluoride and analyzed by gas-liquid chromatography. or cultured at late confluency in the presence of inosine without Cells glucose (17, 18) displayed the presence of regular apical microHT29-Ino villi and tight junctions. In order to determine the possible relationship between the phospholipid pattern and differentiation in the HT29 cell sys Fatty acids16:1 confluent30.8 confluent22.4 colonie5.6 0.8*25.415.80.20.51.70.40.3*0.10.30.2*0.2 9 0.4*38.3 tem, the plasma membrane fractions were prepared from dif 1.530.0 o)718:1 1.927.5 0,918:2 3.03.2 3.820.1 2.32.3 ferent HT29 cell types and human normal colonie cells. Except 1.9*0.8 o)6I8:3o>320:3 1.3r0.7 0.20.4 for the brush border-associated hydrolases, the distribution of 0.10.6 marker enzymes always showed respective similar enhance 0.40.7 0.22.3 0)620:4 0.30.3 0.1HT29-N11. 0.3*0.3 0)620:5 ments of the plasma membrane enzymes and the same decrease + 0.1Normal 0.4Early 0.2Late 0.1 U3HT29-G1C21.340.81.70.51.30.91.3"4.60.30.20.60.30.6 of intracellular enzymes in all plasma membrane fractions as Mean SD of 3 to 5 determinations, expressed as weight percentage of total compared with the corresponding cell homogenates. Thus, it fatty acids. * Ps 0.05 compared with values in HT29-Glc cells (determined by Cochran's can be considered that the degree of purity of the plasma t test. membrane prepared from all cell types was nearly the same with a minor contamination with intracellular organelles. In contrast with these data, the specific activities of the brush border-associated hydrolases in the plasma membrane were PE similarly highly increased as compared with those of the cell fatHnounsaturated Js/Polyunsaturated nj homogenate in both morphologically differentiated HT29 and normal colonie cells. This marked enhancement of alkaline phosphatase and aminopeptidase activities was not found in the plasma membrane isolated from undifferentiated HT29 cells. This distribution of brush border-associated hydrolases related to the differentiation state of HT29 cells is consistent with previous studies (17, 22) as well as with ultrastructural I1IiIii_I_l data reported above. The lipid composition of these purified membranes and their corresponding cell homogenates was investigated in undiffer entiated and differentiated HT29 cells, and in normal colonie cells. Fig. 3. Monounsaturated:polyunsaturated fatty acid ratios in phosphatidylchoChanges in the total or free cholesterol:phospholipid ratio of line and phosphatidylethanolamine. a to c. as in Fig. 2. *, Significantly different cell homogenate, plasma membrane, and various intracellular from respective values found in a with P < 0.01 for phosphatidylcholine and P < 0.05 for phosphatidylethanolamine (determined by Cochran's t test). Columns, membranes occur often in relationship with cellular transfor mean: bars, SD (n = 3 to 5). mation or differentiation. In this study, however, the amounts of free cholesterol and total phospholipid, and consequently the cells. In addition, DMA that originate from plasmalogens were free cholesterol:phospholipid molar ratio in both cell homoge present only in phosphatidylethanolamine with about a 2-fold nate and plasma membrane fraction, did not differ between increased level in normal colonie cells. The differences in fatty undifferentiated and differentiated HT29 cells. A pronounced acid composition of both phosphatidylethanolamine and phos increase of phospholipid content in both cell homogenate and phatidylcholine found between the early confluent and the late plasma membrane that resulted in a decrease in the free cholesconfluent HT29-Ino cells were not due to a depletion of essen terol:phospholipid ratio was found in normal colonie cells. tial fatty acids in the culture medium, since the level of these Recent results indicated that the levels of free cholesterol and
1274

LIPID CHANGES DURING DIFFERENTIATION

OF TUMOR CELLS

total phospholipid were, respectively, slightly or greatly in creased in human colonie tumors as compared with normal intestinal mucosa (47). The reasons for this discrepancy are at this time unknown. In normal colonie cells, comparative phospholipid analyses of the cell homogenate and the plasma membrane showed a high increase in sphingomyelin content and a moderate increase in phosphatidylcholine amount, associated with slight decreases in phosphatidylethanolamine and phosphatidylserine contents and the absence of cardiolipin in the plasma membrane. These differences, notably the enrichment of sphingomyelin, the ab sence of cardiolipin, and the increase of free cholesterol, are in line with the general trend of lipid localization in normal eukaryotic cells (48). These results were also consistent with similarities in the distinct phospholipid composition of each antipodal plasma membrane observed in the rat colonocyte and enterocyte (28, 29, 49, 50), the brush border domain having a specially high sphingomyelin amount and the basolateral do main being enriched to a lesser extent in phosphatidylcholine. The differential phospholipid pattern between the cell ho mogenate and the plasma membrane found in normal colonie cells was also seen in differentiated HT29 cells, with the excep tion of certain variations which were less acute or abolished according to the differentiated cell type. It seems likely that these slight differences in the plasma membrane phospholipid composition between differentiated HT29 and normal colonie cells were, at least in part, dependent on the differentiation state of HT29 cells (16, 19, 20). On the contrary, in undifferentiated HT29 cells, the plasma membrane-phospholipid composition was similar to that of cell homogenate. This organelle had only a moderate increase of the sphingomyelin content and a decrease of cardiolipin. This equalization of the phospholipid pattern between the plasma membrane and the cell homogenate has already been reported in some experimental tumors, notably various rat hepatomas as compared with normal liver (2, 51, 52). This loss of specific phospholipid composition in the plasma membrane was not due to cross-contamination and intermembrane phospholipid redistribution during subcellular fractionation or related to cell growth, but it was dependent on the malignancy of the tumor (52). Moreover, the phospholipid composition of all cell homogenates was similar and, therefore, the biosynthesis of these components was possibly unchanged in different cell types. On the contrary, the phospholipid composition of the plasma mem brane was a function of the differentiation state, suggesting that the polarized distribution of the phospholipids between the exoplasmic and cytoplasmic leaflets of the antipodal domains in epithelial cells (53) was probably altered in undifferentiated HT29 cells, but essentially restored in their enterocytic-differentiated counterparts. It was recently reported that the relative rate of translocation, and not the synthetic step, was the most important factor controlling the phospholipid segregation be tween apical and basolateral membranes in rat renal proximal tubule cells (54). These probable altered sorting and targeting of phospholipids in undifferentiated HT29 cells are also con sistent with the abnormal expression and distribution of other cell surface molecules recently observed in these cells when compared with their differentiated counterparts (21, 23-27). In particular, as already indicated above in this study, the levels of both alkaline phosphatase and /V-aminopeptidase were not in creased in the undifferentiated plasma membrane. The differences in acyl group composition of total or individ

ual phospholipids in relation to cell transformation or differ entiation have often been reported. Tumor cells were usually characterized by a lower proportion of polyunsaturated fatty acids and a higher percentage of monounsaturated fatty acids than the corresponding normal cells (1,5, 55). In agreement with this general tendency, the monounsaturatedrpolyunsaturated fatty acid ratio in both phosphatidylcholine and phosphatidylethanolamine had the highest value in undiffer entiated tumoral HT29 cells but the lowest in normal colonie cells. These changes were essentially the result of an opposite variation in the palmitoleic acid proportion and those of linoleic, arachidonic, and/or linolenic acids in both phospholipids. The differentiation of HT29 cells leads to a partial reversion of the fatty acid pattern characteristic of normal colonie cells. Except for phosphatidylcholine in one differentiated HT29 cell type, the monounsaturated:polyunsaturated fatty acid ratios had the same values. However, this restoration was related to lower changes in the proportion of the fatty acids than the differences found between undifferentiated HT29 cells and nor mal colonie cells. In particular, there was only a decrease in the amount of either arachidonic acid or linoleic and linolenic fatty acids, depending on the differentiated HT29 cell type. The reason for the absence of fatty acid reversion in phosphatidyl choline in differentiated HT29-N cells is unknown but might be due, at least in part, to the low and quite similar proportion of arachidonic acid in this phospholipid in all HT29 cell types as compared with that in phosphatidylethanolamine. These data are consistent with changes in fatty acyl profile observed during in vitro differentiation of various cultured cells (8, 56-59), although controversial data were reported in Friend erythroleukemia cells (7, 57). The fatty acyl changes found in this study did not appear to be due to variations in essential fatty acid uptake related to different nutritional conditions. They might be due rather to alterations in the activities of enzymes respon sible for the transformation of linoleic and linolenic precursors in other polyunsaturated fatty acids (55, 60). The decreased contents of arachidonic acid and both linoleic and linolenic precursors found in undifferentiated HT29 cells might also be related to an increased formation of leukotrienes B4, sulfidopeptide leukotrienes, and prostaglandin E2 as observed in hu man colonie adenocarcinomas when compared with autologous normal colonie mucosa (47). In summary, the results reported here, for the first time to our knowledge, show that the absence of a characteristic phos pholipid plasma membrane composition observed in undiffer entiated HT29 cells reversed essentially to a normal pattern in enterocytic-differentiated counterparts. Moreover, it seems rea sonable to suggest that these significant phospholipid changes might be important in different cellular properties related to malignancy or differentiation in the HT29 experimental model and, in turn, in the human gut. In particular, the phosphatidylcholine:phosphatidylethanolamine ratio is known to be a deter minant in cell-cell interaction, cell-substrate adhesion, endocytosis, receptor binding, and metastatic potential (1,61,62). The level of polyunsaturated fatty acids affects lectin-induced agglutinability, receptor interactions, cytolysis, proliferation, tumorigenicity, and metastatic potential (1, 5, 11, 63, 64). Recently sphingomyelin and more specifically its breakdown products, which are protein kinase C inhibitors and may function as second messengers, appear to be active in signal transduction, receptor activities, phorbol ester-induced responses, action of tumor promoters, and growth factors (65).

1275

LIPID CHANGES DURING DIFFERENTIATION

OF TUMOR CELLS

ACKNOWLEDGMENTS
We are obliged to Dr. A. Zweibaum and Dr. G. Trugnan for the gift of HT29-Ino cells and Dr. S. Champion for critically reading the manuscript. We are very grateful to F. <Cannellini and R. Soirat for the cell culture, M. Roccabianca for the electron microscopy, and J. J. Roccabianca for the photographs.

REFERENCES
1. Schroeder. F., and Gardiner, J. M. Membrane lipids and enzymes of cultured high- and low-metastatic B16 melanoma variants. Cancer Res., 44: 32623269, 1984. 2. Wood, R., Upreti, G. C., and de Antueno, R. J. A comparison of lipids from liver and hepatoma subcellular membranes. Lipids, 21: 292-300, 1986. 3. Calorini, L., Fallani, A., Tombaccini, D., Mugnai, G., and Ruggieri, S. Lipid composition of cultured B16 melanoma cell variants with different lungcolonizing potential. Lipids, 22: 651-656, 1987. 4. Delicoinstantinos, G. Physiological aspects of membrane lipid fluidity in malignancy. Anticancer Res., 7: 1011-1022, 1987. 5. Le Bivic, A., Sari, H., Reynier, M., Lebec, S.. and Bardin, F. Differences in lipid characteristics of autologous human melanoma cell lines with distinct biological properties. J. Nati. Cancer Inst., 79: 1181-1188, 1987. 6. Kaiho, S-I., and Mizuno, K. Inhibition of Friend cell erythrodifferentiation by modification of membrane phospholipid composition by choline analogs. Biochim. Biophys. Acta, 838: 175-178, 1985. 7. Fallani, A., Arcangeli!, A., and Ruggieri. S. Lipid characteristics of Friend erythroleukaemia cells differentiated by dimethylsulfoxide or hexamethylenebisacetamide and of non-inducible clones treated with the inducers. Biochem. J., 252: 917-920, 1988. 8. Ponec, M., Weerheim, A., Kempenaar, J., Mommaas, A-M., and Nugteren, D. H. Lipid composition of cultured human keratinocytes in relation to their differentiation. J. Lipid Res., 29: 949-961, 1988. 9. Connor, J., Bucana, C., Fidler, I. J., and Schroij, A. J. Differentiationdependent expression of phosphatidylserine in mammalian plasma mem branes. Quantitative assessment of outer-leaflet lipid by prothrombinase complex formation. Proc. Nati. Acad. Sci. USA, 86: 3184-3188, 1989. 10. Stubbs, C. D., and Smith. A. D. The modification of mammalian polyunsaturated fatty acid composition in relation to membrane fluidity and function. Biochim. Biophys. Acta, 779: 89-137, 1984. 11. Spector, A. A., and Yorek, M. A. Membrane lipid composition and cellular functions. J. Lipid Res.. 26: 1015-1035. 1985. 12. Boukamp, P., Rupniak, H. T. R., and Fusenig, N. E. Environmental modu lation of the expression of differentiation and malignancy in six human squamous cell carcinoma cell lines. Cancer Res., 45: 5582-5592, 1985. 13. Ponec, M., Weerheim, A., Kempenaar, J., Elias, P. M., and Williams, M. L. Differentiation of cultured human keratinocytes: effect of culture conditions on lipid composition of normal vs. malignant cells. In Vitro Cell. Dev. Biol., 25:689-695, 1989. 14. Roussel, M., Dussaulx, E., Chevalier, G., and Zweibaum, A. Growth-related glycogen levels of human intestine carcinoma cell lines grown in vitro and in vivo in nude mice. J. Nati. Cancer Inst., 65: 885-889, 1980. 15. Pinto, M., Appay, M. D., Simon-Assmann, P., Chevalier, G., Dracopoli, N., Fogh, J., and Zweibaum, A. Enterocytic differentiation of cultured human colon cancer cells by replacement of glucose by galactose in the medium. Biol. Cell., 44: 193-196, 1982. 16. Zweibaum, A., Triadou, N., Kedinger, M.. Augeron, C., Robine-Lon,S., Pinto, M., Roussel, M., and Haffen, K. Sucrase-isomaltase: a marker of foetal and malignant epithelial cells of the human colon. Int. J. Cancer, 32: 407-412, 1983. 17. Wice, B. M., Trugnan, G., Pinto, M.. Rousset, M., Chevalier, G.. Dussaulx, E., Lacroix, B., and Zweibaum, A. The intracellular accumulation of UDP/V-acetylhexosamines is concomitant with the inability of human colon cancer cells to differentiate. J. Biol. Chem., 260: 139-146. 1985. 18. Zweibaum, A. Enterocytic differentiation of cultured human colon cancer cell lines: negative modulation by D-glucose. In: F. Alvarado and C. H. van Os (eds.). Ion Gradient-coupled Transport, pp. 345-353. Amsterdam: Elsevier, 1986. 19. Lacroix, B., Kedinger, M., Simon-Assmann, P., Rousset, M., Zweibaum, A., and Haffen, K. Developmental pattern of brush-border enzymes in the human fetal colon. Correlation with some morphogenetic events. Early Hum. Dev., 9:95-103, 1984. 20. Triadou, N., and Zweibaum, A. Maturation of sucrase-isomaltase complex in human foetal small and large intestine during gestation. Pediatr. Res., 19: 136-138, 1985. 21. Simon-Assmann, P., Bouziges, F., Daviaud, D Haffen, K., and Kedinger. M. Synthesis of glycosaminoglycans by undifferentiated and differentiated HT29 human colonie cancer cells. Cancer Res., 47: 4478-4484, 1987. 22. Trugnan, G., Rousset, M., Chantret, I., Barbt,A., and Zweibaum, A. The posttranslational processing of sucrase-isomaltase in HT29 cells is a function of their state of enterocytic differentiation. J. Cell Biol., 104: 1199-1205, 1987. 23. Ogier-Denis, E., Codogno, P., Chantret, I., and Trugnan, G. The processing of asparagin-linked oligosaccharides in HT29 cells is a function of their state 1276

of enterocytic differentiation. J. Biol. Chem., 263: 6031-6037, 1988. 24. Fantini, J., Martin, J-M., Luis, J., Remy, L., Tirard, A., Marvaldi, J., and Pichn. J. Restricted localization of functional vasoactive inleslinal peplide (VIP) receplors in in vitro differenlialed human colonie adenocarcinoma cells (HT29-D4). Eur. J. Cell. Biol., 46:458-465, 1988. 25. Fantini, J., Rognoni, J-B., Roccabianca, M., Pommier, G., and Marvaldi, J. Suramin inhibits cell growth and glycolytic aclivily and Iriggers differenlialion of human adenocarcinoma cell clone HT29-D4. J. Biol. Chem., 264: 10282-10286, 1989. Fanlini, J., Rognoni, J-B., Culouscou, J-M., Pommier, G., Marvaldi, J., and 26. Tirard, A. Induclion of polarized apical expression and vectorial release of carcinoembryonic antigen (CEA) during ihe process of differenliation of HT29-D4 cells. J. Cell. Physiol., 141: 126-134, 1989. 27. Le Bivic, A., Real, F. X., and Rodriguez-Boulan, E. Vectorial targeling of apical and basolaleral plasma membrane proteins in a human adenocarci noma epithelial cell line. Proc. Nati. Acad. Sci. USA, 86:9313-9317, 1989. 28. Brasitus, T. A., and Schachler, D. Lipid dynamics and lipid-protein interaclions in rat enterocyte basolaleral and microvillus membranes. Biochemistry, 19: 2763-2769, 1980. 29. Brasitus, T. A., and Keresztes, R. S. Prolein-lipid interactions in antipodal plasma membranes of rat colonocytes. Biochim. Biophys. Acta, 77.: 290300, 1984. 30. Brasitus, T. A., Dahiya, R., and Dudeja, P. K. Rat proximal intestinal Golgi membranes: lipid composition and fluidity. Biochim. Biophys. Acta, 958: 218-226, 1988. 31. Fogh, J., and Trempe, G. New human lumor cell lines. In: J. Fogh (ed.), Human Tumor Cells in vitro, pp. 115-141. New York: Plenum Press, 1975. 32. Denis, C., Corlinovis, C., Terrain, B., Viallard, V., Paris, H., and Mural, J. C. Aclivily of enzymes relaled lo carbohydrale metabolism in the HT29 colon adenocarcinoma cell line and tumor. Int. J. Biochem., 16:87-91, 1984. 33. Robine, S., Huet, C., Moll, R., Sahuquillo-Merino, C., Coudrier, E., and Zweibaum, A. Can villin be used to identify malignant and undifferenlialed normal digestive epithelial cells? Proc. Nati. Acad. Sci. USA, 82:8488-8492, 1985. 34. Warren, L. Isolation of plasma membrane from tissue culture-L cells. Melhods Enzymol., 31: 156-162, 1974. 35. Perdue, J. F. The isolation and characlerization of plasma membrane from cultured chicken embryo fibroblasls. Methods Enzymol., 31: 162-168, 1974. 36. Kessler, M., Acuto, O.. Storelli, C., Murer, H., Muller, M., and Semenza, G. A modified procedure for rapid preparation of efficiently transporting vesicles from small intestinal brush-border membrane. Biochim. Biophys. Acta, 506: 136-154, 1978. 37. Rawyler, A. J., Roelofsen, B., Op Den Kamp, J. A. F., and Van Deenen, L. L. M. Isolation and characterization of plasma membranes from Friend erylhroleukaemic cells. A sludy wilh sphingomyelinase C. Biochim. Biophys. Acia, 730: 130-138, 1983. 38. Gralecos, D., Knibiehler, M., Benoil, V., and Semeriva, M. Plasma mem brane from ral inlestinal epithelial cells of differenl slages of maturation. I. Preparation and characterization of plasma membrane subfractions originaling from crypl cells and from villous cells. Biochim. Biophys. Acta, 5/2: 508-524, 1978. 39. Ames, B. N. Assay of inorganic phosphate, lotal phosphate, and phosphalases. Melhods Enzymol., 8: 115-118, 1966. 40. Louvard, D., Maroux, S., Baratti, J., Desnuelle, P., and Mutaflschiev, S. On the preparation and some properlies of closed membrane vesicles from dog duodenal and jejunal brush border. Biochim. Biophys. Acia, 297: 747-763, 1973. 41. Morton, R. K. The purification of alkaline phosphalases of animal tissues. Biochem. J., 57: 595-603, 1954. 42. Sotlocosa, G. L., Kuylenslierna, B., Ernster, L., and Bergstrand, A. An electron transport system associated with the outer membrane of liver milochondria. A biochemical and morphological sludy. J. Cell. Biol., 32: 415438. 1967. 43. Cooperstein, S., and Lazarow, A. A microspectropholomelric melhod for ihe di-ter minal ion of cytochrome oxidase. J. Biol. Chem., 189: 665-670, 1951. 44. Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. Protein measurements wilh Ihe Folin phenol reagent. J. Biol. Chem., 193: 265-275, 1951. 45. Folch, J., Lees, M., and Sloane-Stanley, G. H. A simple melhod for ihe isolalion and purification of total lipids from animal tissues. J. Biol. Chem., 226:497-509,1957. 46. Varkovsky, V. E., and Kostetsky. E. Y. Modified spray for Ihe delection of phospholipids on thin-layer chromatograms. J. Lipid Res., 9:396-402,1968. 47. Berstein, L. M., Kovaleva, I. G., Pravosudov, I. V., and Dilman, V. M. Pecularities of hormonal regulation of adenylale cyclase and lipid composi tion in colonie and rectal tumors. Neoplasma, 36: 465-470, 1989. 48. Dawidowicz, E. A. Dynamics of membrane lipid metabolism and turnover. Annu. Rev. Biochem., 56:43-61, 1987. 49. Brasilus, T. A., and Dudeja, P. K. Regional differences in Ihe lipid composilion and fluidily of rat colonie brush-border membranes. Biochim. Biophys. Acia, 819: 10-17, 1985. 50. Brasilus, T. A., and Schachler, D. Lipid composition and fluidity of rat enterocyte basolateral membranes. Regional differences. Biochim. Biophys. Acta, 77*: 138-146, 1984. 51. Bergelson, L. D., Dyatloviiskaya, E. V., Torkhovskaya, T. I., Sorokina, I. B., and Gorkova, N. P. Phospholipid composilion of membranes in ihe tumor

LIPID CHANGES DURING DIFFERENTIATION

OF TUMOR CELLS

cells. Biochim. Biophys. Acta, 210: 287-298, 1970. 52. Bergeisen, L. D., Dyatlovitskaya, E. V., Sorokina, I. B., and Gorkova, N. P. Phospholipid composition of mitochondria and microsomes from regener ating rat liver and hepatomas of different growth rate. Biochim. Biophys. Acta, 360: 361-365, 1974. 53. Simons, K., and Van Meer, G. Lipid sorting in epithelial cells. Biochemistry, 27:6197-6202, 1988. 54. Molitoris, B. A., and Simon, F. R. Maintenance of epithelial surface mem brane lipid polarity: a role for differing phospholipid translocation rates. J. Membr. Biol., 94: 47-53, 1986. 55. Robert, J., Montaudon, D., and Hugues, P. Incorporation and metabolism of exogenous fatty acids by cultured normal and tumoral glial cells. Biochim. Biophys. Acta, 752: 383-395, 1983. 56. Robert, J., Mandel, P., and Rebel, G. Membrane lipids in bromodeoxyuridine-differentiated astroglial cells in culture. Lipids, 14: 852-859, 1979. 57. Zwingelstein, G., Tapiero, H., Portoukalian, J., and Fourcade, A. The effect of dimethylsulfoxide on the lipid composition of inducible and non-inducible Friend leukemia cells. Biochem. Biophys. Res. Commun., 108: 437-446, 1982. 58. Lampe, M. A., Williams, M. L., and Elias, P. M. Human epidermal lipids:

59. 60. 61. 62. 63.

characterization and modulation during differentiation. J. Lipid Res., 24: 131-140, 1983. Isseroff, R. R., Ziboh, V. A., Chapkin, R. S., and Martinez, D. T. Conversion of linoleic acid into arachidonic acid by cultured murine and human keratinocytes. J. Lipid Res., 2: 1342-1349, 1987. Dunbar, L. M., and Bailey, J. M. Enzyme deletions and essential fatty acid metabolism in cultured cells. J. Biol. Chem., 250: 1152-1153, 1975. Kier, A. B., and Schroeder, F. Development of metastatic tumors in athymic (nude) mice from LM cells grown in vitro. Transplantation, 33: 274-279, 1982. Schroeder, F. Altered phospholipid composition affects endocytosis in cul tured LM fibroblasts. Biochim. Biophys. Acta, 649: 162-174, 1982. Stubbs, C. D., and Smith, A.'D. The modification of mammalian membrane

polyunsaturated fatty acid composition in relation to membrane fluidity and function. Biochim. Biophys. Acta, 779: 89-137, 1984. 64. El Battari, A., Ah-Kye, E., Muller, J.-M., Sari, H., and Marvaldi, J. Modi fication of HT29 cell response to the vasoactive intestinal peptide (VIP) by membrane fluidization. Biochimie, 67:1217-1224, 1985. 65. Hannun, Y. A., and Bell, R. M. Functions of sphingolipids and sphingolipid breakdown products in cellular regulation. Science (Wash. DC), 243: 500507, 1989.

1277