Regulation of cholesterol synthesis and binding of lipoproteins in cultured rat intestinal epithelial cells

12

Riochimica

et Biophysics

Actu 816 (1986) 12-79

Elsevier

BBA 52171

Regulation of cholesterol synthesis and binding of lipoproteins in cultured rat intestinal epithelial cells

Jean-Louis y Centre

Ronald Barbaras Patrick Rampal S*

d’Hipato-Gastro-Entkoologie,

Hspital

LP7300,

Universitk

de Cimiez, BP 179, 06003 Nice Cedex, and h Centre de Biochimie

Cholesterol

synthesis

du CNRS,

de Nice, Pare Valrose, 06034 Nice Cedex (France)

(Received

Key words:

NCgrel b and

July 17th. 1985)

regulation; Hydroxymethylglutaryl-CoA Lipoprotein binding; (Rat intestinal

reductase; cell)

Intestinal

cholesterogenesis;

The regulation of cholesterol synthesis has been studied using a rat epithelial intestinal cell line (IRD 98) as a cellular model. As observed in other cell types, mevinolin increases the levels of 3-hydroxy-3-methylglutaryl coenzyme A reductase (EC 1.1.1.34) and concomitantly reduces the incorporation of [ i4Clacetate into cholesterol. Free cholesterol is able to suppress reductase activity. In contrast, when cells are shifted from standard culture medium to lipoprotein-deficient medium, an increase in hydroxymethylglutaryl-CoA reductase specific activity (24fold) is observed. The possible regulatory roles of the different classes of human lipoproteins were thus compared. The effects of a long-term exposure to LDL and HDL vary according to cell density. In actively growing cells, VLDL and LDL cause a decrease in the level of hydroxymethylglutaryl-CoA reductase, whereas HDL do not have a significant effect. In contrast, in subconfluent preresting cells, HDL provoke large decreases in hydroxymethylglutaryl-CoA reductase activity as compared to VLDL and LDL. While LDL binding is constant, the maximal binding capacity of HDL in subconfluent cells is seven times that of actively growing cells. Altogether, these results suggest an important role for HDL in the regulation of intestinal cholesterol synthesis.

Introduction The intestinal mucosa is a major site for the synthesis of endogenous cholesterol [l-3]. In the villous cells, de novo synthetized cholesterol, in addition to dietary and biliary cholesterol, contributes to the formation of chylomicrons, VLDL and HDL [4]. Cholesterol is also essential for membrane biogenesis during growth and maturation of crypt cells [5]. The regulation of cholesterogenesis in the intestinal mucosa is not as well documented as that in the liver and published results are often

* To whom correspondence

OOOS-2760/86/$03.50

should be addressed.

0 1986 Elsevier

Science Publishers

conflicting [6-81. Nevertheless, variations in cholesterol synthesis observed in rat small intestine, according to a circadian rhythm [9], have been well correlated with hydroxymethylglutarylCoA reductase activity [lo]. As in other tissues, this enzyme is rate-limiting for cholesterol synthesis; it is found in the microsomal fraction of crypt and villous intestinal cells [ll] and is feed-back regulated. Studies carried out in vivo in various rat extrahepatic tissues have shown that intestinal sterol synthesis was stimulated in animals rendered hypolipoproteinemic after injection of 4aminopyrazolo[3,4-dlpyrimidine and further suppressed after infusion of LDL [6]. On the other

B.V. (Biomedical

Division)

73

hand, both LDL and HDL have been reported to be catabolized at the intestinal level in the rat [12,13]. More recently, using organ cultures of rabbit intestinal mucosa, Stange et al. confirmed the regulatory role of LDL and made the proposal that VLDL and HDL are also potent regulators of intestinal cholesterol synthesis [8]. That HDL might play a significant role in intestinal cholesterol metabolism is further supported by the characterization of separate binding sites for LDL and HDL in isolated rat intestinal cells [14]. The recent establishment in our laboratory of an intestinal epithelial cell line from rat fetus (IRD 98; Ref. 15) thus offered a good opportunity to study the regulation of cholesterol synthesis in vitro.

the method of Lowry et al. [18] using bovine serum albumin as standard; phospholipids, triacylglycer01, free and esterified cholesterol were determined by using commercial kits from Biolyon and Boehringer. Radioiodination of LDL and HDL was carried out according to the iodine monochloride method [19]. After iodination, the labeled lipoproteins were dialysed against phosphate-buffered saline and sterilized by filtration. After extraction with chloroform/methanol, less than 5% of the radioactivity of each lipoprotein class was found to be associated with the lipid fraction. The specific radioactivities of preparations used in this study were between 50 and 100 cpm per ng of protein.

Materials and Methods

Incorporation of radioactive precursors into cholesterol and assay of hydroxymethylglutatyl-CoA reductase activity. IRD 98 cells were maintained in

Cell culture. IRD 98 cells were grown at 37°C in a humidified 5% CO, atmosphere in the presence of Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, penicillin (50 U/ml) and streptomycin (50 pg/ml) as previously described [15] and defined as standard medium. Media were changed three times weekly. Cell number was determined by using a Coulter counter. Cells were plated at 2.5 . lo3 cells per cm2 either in 60 mm plastic Petri dishes for hydroxymethylglutaryl-CoA reductase assay, [‘4C]acetate and [‘4C]mevalonate incorporation, or in 4 cm2 multi-well plates for binding experiments. isolation and labeling of lipoproteins. Human lipoprotein classes were separated by sequential ultracentrifugation [16] of pooled plasma from normal donors. Isolated fractions were stored in the presence of NaN, (1 mg/ml) and EDTA (0.1 mg/ml) at 4°C and dialysed exhaustively against phosphate-buffered saline supplemented with Ca” and Mg 21 before utilization. Lipoprotein-deficient fetal bovine serum was prepared by ultracentrifugation (150000 X g 70 h) at 16°C after adjustment of density to 1.22 g/ml by adding solid KBr. The infranatant was exhaustively dialysed against saline. The purity of the different lipoproteins was assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis [17]. Standard methods were used to determine the composition of lipoprotein classes; protein content was determined by

the presence of 2 FCi of [l-‘4C]acetate or of [2-‘4C]mevalonate for I6 h at 37°C under various culture conditions as indicated in the figure legends. Washed cells were solubilized in 0.1 N NaOH, and cellular lipids extracted after acidification according to Bligh and Dyer [20]. Cholesterol was separated from other lipid classes by thin-layer chromatography using benzene/ethyl acetate/ formic acid (80 : 20 : 1) as developing solvent. The areas corresponding to cholesterol (run in parallel as a standard and detected with iodine vapours) were scraped and counted by liquid scintillation. Hydroxymethylglutaryl-CoA reductase assays were carried out according to Alberts et al. 1211 on activated cell extracts previously prepared according to the same authors. Binding assays. IRD 98 cells were grown and treated as indicated in the legend of Fig. 4. After washing with ice-cold lipoprotein-deficient medium, cells were maintained in this medium at 4°C for 15 min before any addition. The cells were then incubated at 4’C with the labeled LDL or HDL (O-100 yg protein/ml) in the absence (total binding) or in the presence (nonspecific binding) of 800 pg protein/ml of the corresponding unlabeled lipoproteins. At the end of the incubation (2 h for saturation experiments), the medium was removed and cells were rapidly washed at 4°C four times with 50 mM Tris-HCl buffer (pH 7.4) containing 0.15 M NaCl and 2 mg/ml albumin

74

and three times with isotonic Tris-HCl buffer alone. After solubilization in 0.5 ml of 0.1 N NaOH, the cell digest was assayed for radioactivity and for protein content. The specific binding was determined by subtraction of the amount of nonspecifically bound lipoprotein from total binding. The percentages of nonspecific binding of LDL and HDL were 27 and 22%, respectively. The ratios of bound to free ‘251-labeled lipoprotein were plotted against bound lipoprotein according to the method of Scatchard [22]. Materials. Culture products including medium, antibiotics and serum were purchased from Gibco. Cholesterol, 3-hydroxy-3-methylglutaryl coenzyme A, NADPH and bovine serum albumin were from Sigma Chem. Co. Mevinolin was kindly provided by Dr. A.W. Alberts (Merck Sharp and Dohme Research Laboratories, Rahway, NJ, U.S.A.) and human plasma by Dr. R. Follana (Centre de Transfusion Sanguine, Nice, France). 3-Hydroxy3-methyl[3-i4C]glutaryl coenzyme A and DL-[2“C]mevalonic acid dibenzylethylenediamine salt were purchased from Amersham. [ l-l4 C]Acetic acid and Na’251 were products from ‘Commissariat a 1’Energie Atomique’.

TABLE 1 EFFECT OF LIPOPROTEIN-DEFICIENT MEDIUM ON CHOLESTEROL METABOLISM IRD 98 cells were plated and grown for 2 days in standard medium. Cells were then incubated with either fresh standard medium or lipoprotein-deficient medium. Incorporations of [t4C]acetate, [ “C]mevalonate into cholesterol and assays of hydroxymethylglutaryl-CoA (HMG-CoA) reductase activity were measured 16 h later according to the procedures described in Materials and Methods. The data are the mean values of triplicate incubations in two separate experiments + SD. Culture conditions

Standard medium Lipoproteindeficient medium

HMG-CoA

[l-‘4C]Acetate incorporation

[2-‘4C]Mevalonate incorporation

reductase (pmol/min

(dpm/mg protein)

(dpm/mg protein)

per mg protein)

30000*2500

10000~800

lOkO.7

65OOOk4000

11000~700

2051.2

Results Experiments presented in Table I and Figs. 1 and 2 show that cholesterol synthesis in IRD 98 intestinal cells is under the control of exogenous cholesterol and that hydroxymethylglutaryl-CoA reductase is the rate-controlling feed-back regulated enzyme of this pathway. Cells previously maintained in lipoprotein-deficient medium exhibit a 2-fold increase in [14C]acetate incorporation into cholesterol and a concomitant rise in the level of total hydroxymethylglutaryl-CoA reductase activity when compared to control cells continuously maintained in standard medium (Table I). In contrast, [‘4C]mevalonate incorporation into cholesterol does not change, whatever the culture conditions. Increasing concentrations of free cholesterol, added as an ethanolic solution to cells exposed to lipoprotein-deficient medium (Fig. 1) leads to a clear dose-dependent reduction in hydroxymethylglutaryl-CoA reductase activity. A cholesterol concentration as low as 12.5 PM brings within 16 h a 70% decrease in reductase activity.

.1LLY 1 : CHOLESTERoLfTu MDI”M5hlm) Fig. 1. Effect of cholesterol addition on hydroxymethylglutarylCoA (HMG-CoA) reductase activity. Cells were previously grown for 4 days in standard medium and then shifted to lipoprotein-deficient medium. 24 h later, cells were fed with fresh lipoprotein-deficient medium supplemented with various concentrations of cholesterol as indicated. Cholesterol was added as an ethanolic solution, the final concentration of ethanol being kept constant at 1%. Control cells received 1% ethanol. 16 h later, cells were washed and harvested for hydroxymethylglutaryl-CoA reductase assays. Each value is the mean of duplicate determinations from duplicate dishes which did not differ by more than 8% in a single experiment.

75

0-a s =

0

om LPOPROTEN

10

Km

Km

MEVINOLIN

1

I

02

0.1

CHOLESTEROL

N

h4ElXJM

(d

m

(no)

Fig. 2. Effect of mevinolin on hydroxymethylglutaryl-CoA (HMG-CoA) reductase activity and [14C]acetate incorporation into cholesterol. 4 days after inoculation, cells were fed with standard medium supplemented with various concentrations of mevinolin as indicated. One set of dishes received, in addition, 2 PCi of [14C]acetate. [‘4C]Acetate incorporation into cholesterol and hydroxymethylglutaryl-CoA reductase activity were determined 16 h later as described in Materials and Methods. Each point is the mean of duplicate determinations from duplicate dishes in a single experiment. Variations did not exceed f 10%.

Furthermore, treatment of IRD 98 cells with mevinolin (Fig. 2) a competitive inhibitor of hydroxymethylglutaryl-CoA reductase, dramatically reduces [14C]acetate incorporation into cholesterol (EC,, = 50 nM) and concomitantly increases hydroxymethylglutaryl-CoA reductase specific activity (15fold at 5 PM mevinolin). Similar results have been reported using other cell types [23,24]. The biological response observed both on [14C]acetate incorporation into cholesterol and on hydroxymethylglutaryl-CoA reductase activity for IRD 98 cells maintained in lipoprotein-deficient medium led us to examine the possible influence of human lipoproteins on this latter parameter. These experiments (Fig. 3) were carried out with actively dividing cultures (low density = 5-10 . lo3 cells/cm’; Fig. 3A) and with subconfluent culture (high density = 6-10. lo4 cells/cm2; Fig. 3B), since (i) cell density has already been reported to influence hydroxymethylglutaryl-CoA reductase

LlPOPROTElN

CHOLESTEROL

N MELXJ’A

hd

Fig. 3. Effect of human lipoproteins on hydroxymethylglutarylCoA (HMG-CoA) reductase specific activity. On day 2 (A) or 6 (B) after inoculation, cells were fed for 24 h with lipoproteindeficient medium. Cells were further incubated for 16 h in fresh lipoprotein-deficient medium containing the indicated concentrations of VLDL (A), LDL (0) or HDL (0). Cell extracts were then prepared for assay of hydroxymethylglutaryl-CoA reductase activity. Each point corresponds to the mean of duplicate determinations from duplicate dishes whose values did not differ by more than 8% in three different experiments. T!Tand arrow indicate the specific activity of hydroxymethylglutaryl-CoA reductase measured in cells continuously maintained in standard medium.

activity [25] as well as binding of lipoproteins [26] on various cell types and (ii) IRD 98 cells have been previously shown to express some differentiated functions of mature intestinal cells when reaching confluence [15]. Fig. 3 first shows that, as expected [25], the level of hydroxymethylglutarylCoA reductase activity in cells previously exposed to lipoprotein-deficient medium and further maintained for 16 h in the absence of any added lipoprotein is higher for low-density than for high-density cultures. Considering low-density cultures (Fig. 3A), it is clearly apparent that VLDL as well as LDL added for 16 h into lipoprotein-deficient medium are able

76

to decrease the level of hydroxymethylglutaryl-CoA reductase activity. The decrease in hydroxymethylglutaryl-CoA reductase specific activity is to 60% of control at and above 25 PM lipoprotein cholesterol. Under the same conditions, HDL have hardly any effect, even when present at a concentration of 200 I_LMlipoprotein cholesterol. In contrast, on high-density cultures (Fig. 3B), HDL leads to a 50% suppression of hydroxymethylglutaryl-CoA reductase activity at 25 FM lipoprotein cholesterol, whereas VLDL and LDL only lead to 20-30% decrease up to 200 PM. In both cases, the maximal decrease in hydroxymethylglutaryl-CoA reductase activity, induced by the most effective class of lipoprotein, leads to a level which is similar to that observed for cultures maintained in standard medium. These results, showing a differential regulatory effect of HDL cholesterol according to cell density, suggested some modifications affecting properties of lipoprotein surface receptors.

Following this hypothesis, we further studied the binding of iodinated human LDL and HDL on IRD 98 cells under the same previous culture conditions. These experiments were carried out at 4°C in order to minimize the internalization and degradation processes. Under these conditions, time-course experiments have shown that specific binding (which accounted for at least 73% of total binding) for both classes of lipoprotein at both cell densities reached plateau values after 2 h of incubation (not shown). The specific binding data obtained according to this methodology with increasing concentrations of LDL and HDL are presented in Fig. 4A and B, for low- and high-density cultures, respectively. In each case, saturation curves are obtained and the Scatchard plot analysis shows the existence of a single class of binding sites both for LDL and HDL. The binding parameters of “‘I-labeled LDL do not change notably when comparing actively dividing with subconfluent cells; K, values of 10.8 and 12 pg/ml and

r

1251

Lipoprotein

-_

1

50

(&ml)

125

I

Lipoprotein

0

(j.a/ml)

Fig. 4. Concentration dependence of human ‘251-labekd LDL and HDL binding to IRD 98 cells at low and high cell density. After 1 day (A, low celt density = 5-10.103 cells/cm2) or 3 days of growth (B, high density = 6-lo-lo4 cells/cm2) in standard medium, ceils were fed with lipoprotein-deficient medium for 24 h. Indicated concentrations of human ‘251-labeled LDL (0) or human ‘25i-labeIed HDL (0) expressed in pg protein/ml were added for 2 h at 4°C in the absence (total binding) or in the presence of 800 pg/ml of the respective unlabeled lipoproteins (nonspecific binding). Each point represents the specific binding value calculated as described in Materials and Methods from the mean of triplicate dishes. The variability between values from triplicate dishes did not exceed 8% in two experiments for low density cultures (A) and in three experiments for high density cultures (B). Scatchard plots of lipoprotein binding are shown in the insets (same symbols).

a maximal binding capacity of 410 and 380 ng/mg cell protein are obtained for low- and high-density cultures, respectively. Under both conditions the calculated maximal number of LDL binding sites per cell was about 4.5. 105.In contrast, a 7-fold increase in maximum binding of ‘251-labeled HDL was clearly observed between low- and high-density cultures from 7.6. lo5 to 5.3. 10h binding sites per cell with K, values of 5 and 16.8 pg/ml, respectively. Discussion The present data obtained on cultured intestinal cells are, to our knowledge, the first example of a differential regulation of hydroxymethylglutaryl-CoA reductase by LDL and HDL with reference to the proliferation state. With respect to the regulation of cholesterol synthesis by human LDL, the behavior of actively growing IRD 98 cells is similar to that already reported in other cultured cell types from various species [27-301. The decreased level of hydroxymethylglutaryl-CoA reductase is observed at concentrations around 25 PM lipoprotein cholesterol. This diminution is more pronounced with actively growing than with subconfluent cells, suggesting that regulation of hydroxymethylglutaryl-CoA reductase by LDL is impaired at high cell density, as observed in bovine vascular endothelial cells [27]. In contrast to this situation, human HDL have no significant effect on the reductase in actively growing IRD 98 cells even at a concentration as high as 200 PM lipoprotein cholesterol but are able to exert a potent suppressive effect on the enzyme in subconfluent cells at a concentration as low as 25 PM lipoprotein cholesterol. The few data concerning the influence of HDL on the level of hydroxymethylglutaryl-CoA reductase on various cell types [8,28,29,31,32] have shown either an absence of effect or an increase in hydroxymethylglutaryl-CoA reductase activity and/ or in [i4C]acetate incorporation into cholesterol. In normal human fibroblasts, HDL, suppress, while HDL, stimulate hydroxymethylglutaryl-CoA reductase activity [33]. It must be emphasized, however, that at high concentrations HDL, have also been shown to progressively suppress hydroxymethylglutaryl-CoA reductase activ-

ity to basal levels in virus-transformed human lymphoblastoid cells [34]. In vitro studies dealing with the regulation of cholesterogenesis in intestinal mucosa, previously conducted with organotypic cultures, have given contradictory results [7,8]. After a short-term incubation (6 h), dog lipoprotein subclasses failed to affect intestinal cholesterol synthesis in canine intestinal mucosa in the concentration range 0.2-0.8 mM lipoprotein cholesterol [7]. In contrast, addition of rabbit LDL for 24 h resulted in a dose-dependent suppression of reductase activity between 0.2 and 1.0 mM lipoprotein cholesterol, whereas a biphasic response to VLDL and HDL with a 2-3-fold stimulation at 0.1-0.3 mM and suppression at higher concentrations was observed in cultured rabbit intestinal mucosa [8]. The failure to establish any lipoprotein effect in the first study is likely to be due to the shortness of incubations. On the other hand, the discrepancies between our results and those reported with rabbit organ cultures can be easily explained by the differences in the origin of the biological materials and the culture procedure. The characteristics of binding of human lipoproteins on IRD 98 cells are in good agreement with those determined for both human and rat lipoproteins on isolated rat intestinal mucosal cells by Suzuki et al. [14]. The behavior of rat epithelial intestinal cells thus seems particular as compared to other rat tissues which have been shown to bind human LDL poorly [35-371. Interestingly, the emergence of the biological effect of HDL on cholesterol synthesis in subconfluent IRD 98 cells is correlated with an increase in the maximum binding capacity for this class of lipoprotein under the same conditions. It is important to note that, whereas a 7-fold increase in HDL binding capacity occurs in subconfluent as compared to sparse IRD 98 cells, LDL binding remains unchanged. These results allow us to exclude the possibility of some loss of the apolipoprotein B/E receptor, in contrast to the situation observed in adult liver [38,39], and strongly support the development of distinct binding sites for HDL in subconfluent IRD 98 cells as already described in other cell types [30,38-421. Furthermore, these observations are also in agreement with recent studies which demonstrate the existence of a specific binding of

78

human apolipoprotein E-free HDL to the basolateral plasma membrane of isolated rat intestinal mucosal cells [43]. Nevertheless, the nature of the apolipoprotein, if any, involved in the recognition of human HDL at the cell surface remains to be established [38,39,43,44]. The increase in the number of HDL binding sites during growth of IRD 9X cells is likely to reflect some maturation of the plasma membrane as a function of time in culture before the acquisition of differentiated properties associated with the apical plasma membrane [15]. Since LDL uptake and clearance in rat intestine are known to decrease towards the villus tip [45], it is attractive to think that HDL might play a critical role in the regulation of cholesterogenesis in differentiated intestinal epithelial cells. Acknowledgements

The authors are grateful to Professors G. Ailhaud and M.C. Carey for helpful discussions, and Professor R. Green for careful reading of the manuscript. They also wish to thank Mrs. G. Oillaux for her expert secretarial assistance. This work was supported by grants from Institut National de la Sante et de la Recherche Medicale (ATP 60.‘78.92), from Fondation pour la Recherche Medicale and from Faculte de Medecine de Nice. References 1 Srere. P.A.. Chaikoff, I.L.. Treitman, S.S. and Burstein, L.S. (1950) 3. Biol. Chem. 182, 629-634 2 Dietschy, J.M. and Siperstein, M.D. (1967) J. Lipid Res. 8, 977104 3 Dietschy, J.M. and Wilson, J.D. (1968) J. Clin. invest. 47, 166-174 4 Green. P.H.R. and Glickman. R.M. (1981) 3. Lipid Res. 22, 115331173 5 Stange. E.F.. Preclik, G., Schneider, A., Alavi, M. and Ditschuneit. H. (1981) B&him. Biophys. Acta 663,613-620 6 Andersen. J.M. and Dietschy. J.M. (1977) J. Biol. Chem. 252. 3652-3659 7 Gehhard, R.L. and Cooper. A.D. (1978) J. Biol. Chem. 253, 2790-2796 8 Stange. E.F., Alavi. M.. Schneider, A., Preclik, G. and Ditschuneit, H. (1980) Biochim. Biophys. Acta 620. 520-527 9 Edwards, P.A., Muroya, H. and Gould, R.G. (1972) J. Lipid Res. 13, 396-401 10 Shefer. S.. Hauser, S., Lapar, V. and Mosbach, E.H. (1972) J. Lipid Res. 13, 571-573

11 Merchant, J.L. and Heller, R.A. (1977) J. Lipid Res. 18, 122-133 12 Roheim, P.S., Rachmiiewitz, D., Stein, 0. and Stein, Y. (1971) Biochim. Biophys. Acta 248. 315-329 13 Stein, Y.. Halperin, G. and Stein, 0. (1981) B&him. Biophys. Acta 663, 569-514 14 Suzuki, N., Fidge, N., Nestel, P. and Yin, J. (1983) J. Lipid Res. 24, 253-264 15 Negrel, R., Rampal, P., Nano, J.L.. Cavenel, C. and Ailhaud, G. (1983) Exp. Cell Res. 143, 427-437 16 HaveI, R.J., Eder, H.A. and Bragdon, J.H. (1955) J. Clin. Invest. 34, 1345--1353 17 Laemmli, U.K. (1970) Nature 227, 680-685 18 Lowry, O.H.. Rosebrough, N.J.. Farr. A.L. and Randall, R.J. (1951) J. Biol. Chem. 193. 265-275 19 Bilheimer, P.W., Eisenberg, S. and Levy. R.I. (1972) Biochim. Biophys. Acta 260, 212-221 20 Bligh. E.G. and Dyer, W.J. (1959) Can. J. Biochem. Physiol. 37,911-917 21 Alberts. A.W., Chen, C., Kuron, G., Hunt. V., Hoffman, C.. Rothrock, J., Lopez, M.. Joshua, H., Harris, E., Patchett, A.. Monaghan, R.. Currie. S., Stapley, E.. Alberts-Schonberg, G., Hensens. 0.. Hirshfield, J.. Hoogsteen, K., Liesch, J. and Spinger, J. (1980) Proc. Nat]. Acad. Sci. USA 77, 3957-3961 22 Scatchard, G. (1949) Ann. N.Y. Acad. Sci. 51, 660-672 23 Brown, M.S.. Faust. J.R., Goldstein, J.L., Kaneko, I. and Endo, A. (1978) J. Biol. Chem. 253, 1121-1128 24 Kita. T.. Brown, MS. and Goldstein, J.L. (1980) J. Clin. Invest. 66. 1094-1100 25 Chen, H.W. (1981) J. Cell. Physiol. 108, 91-97 26 Kenagy, R., Bierman. E.L. and Schwartz, S. (1983) J. Cell. Physiol. 116, 404-408 21 Fielding, P.E.. Vlodavski, I., Gospodarowicz. D. and Fielding C.J. (1979) J. Biol. Chem. 254, 749-755 28 Brown, M.S., Dana, S.E. and Goldstein, J.L. (1973) Proc. Natl. Acad. Sci. USA 70. 2162-2166 G., 29 Stange. E.F., Fleig, W.E.. Schneider, A., Nether-Fleig, Alavi, M., Preclik. G. and Ditschuneit. H. (1982) Atherosclerosis 41. 67-80 R.. Grimaldi. P., NCgrel, R. and Ailhaud, G. 30 Barbaras. (1985) Biochim. Biophys. Acta 845, 492-501 D. (1982) 31 Cohen, D.C., Massoglia, S.L. and Gospodarowicz, J. Biol. Chem. 257. 9429-9437 32 Gai, D., MacDonald. P.C.. Porter. J.C. and Simpson, E.R. (1982) Arch. Biochem. Biophys. 214, 726-733 33 Daer. W.H., Gianturco, S.H.. Patsch, J.R., Smith, L.C. and Gotto, A.M., Jr. (1980) Biochim. Biophys. Acta 619, 287-301 34 Leikin, A.I., Mikovilovic, M. and Scanu, A.M. (1982) J. Biol. Chem. 257,14280-14287 T.L., Pitas, R.E. and Mahley, R.W. (1980) J. 35 Innerarity, Biol. Chem. 255, 11163-11172 S.H. and Steinberg, 36 Drevon, C.A., Attie, A.D., Pangburn, D. (1981) J. Lipid Res. 22, 37-46 37 Koelz, H.R., Sherill, B.C.. Turley. SD. and Dietschy, J.M. (19X2) J. Biol. Chem. 257, 8061-8072 T.L. and Mahley, R.W. (1981) J. 38 Hui, D.Y., Innerarity, Biol. Chem. 256, 56465655

79

39 Mahley. R.W., Hui. D.Y., Innerarity, T.L. and Weisgraber, K.H. (1981) J. Clin. Invest. 68, 1197-1206 40 Biesbroeck, R., Oram. J.F., Albers, J.J. and Bierman, E.L. (1983) J. Clin. Invest. 71, 525-539 41 Rifici, V.A. and Eder, H.A. (1984) J. Biol. Chem. 259, 13814-13818 42 Fidge, N.H., Nestel, P.J. and Suzuki, N. (1983) Biochim. Biophys. Acta 753. 14-21

43 Kagami, A.. Fidge, N.H., Suzuki, N. and Nestel, P.J. (1984) Biochim. Biophys. Acta 795, 179-190 44 Tabas. I. and Tall. A.R. (1984) J. Biol. Chem. 259. 13897-13905 45 Stange. E.F. and Dietschy, J.M. (1983) Proc. Natl. Acad. Sci. USA 80. 5739-5743