- Email: [email protected]
Cholesterol regulates high-density lipoprotein interaction with isolated epithelial cells of human small intestine
266
Biochimicu
et Blophysica Am
919 (1981) 266-274 Elsevier
BBA 52540
Cholesterol regulates high-density lipoprotein interaction with isolated epithelial cells of human small intestine D.D. Sviridov a, I.G. Safonova a, V.P. Tsybulsky a, A.G. Talalaev ‘, S.N. Preobrazensky a, V.S. Repin a and V.N. Srnirnov a r?~nstif~te of ExFe~i~e~ta~ Cardioluw, and 6 Pathology Depariment, 1st Moscow
(Revised
Key words:
USSR
Card&lo@ Research Center
Chiidren‘s Hospital, ~Uoseow (U.S. S. R.)
(Received 12 June 1986) manuscript received 25 February
HDL binding;
Cholesterol;
Lipoprotein;
1987)
(Epithelial
cell)
The effect of choksterol on the interaction of ~gh-densi~ lipoprotein (HDL) with isolated haan smah intestine epitbelial cells (enterocytes) was studied. ‘251-labeled HDL, binding by these cells was enhanced in response to cholesterol loading of the cells in a time- and dose- dependent manner. Preincubation of the cells with cholesterol led to the enhancement both of the number of binding sites and the binding affinity. The enhancement of binding correiated with the cellular cholesterol content. Cycloheximide (0.5 mM) inhibited uptahe of cholesterol by enterocytes and’bloched its effect on ‘251-labeled HDL, binding. The effect of cholesterol on ‘251-labeled HDL, degradation had a double-phase character. At concentrations lo-20 pg/ml, the degradation rate was rapidly elevated, but further increase in cholesterol ~oneen~tion led to a fail in the degradation rate. Incubation of enteroqtes with HDL, resulted in the efflux of cholesterol from cells and its incorporation into HDL,. The results obtained mahe it possible to assume that binding and degradation of ‘251-labeled HDL, by human enterocytes are independently regulated by the eel1 total cholesteroI content. Binding of HDL by enterocytes may result both in the degradation of HDL and cholesterol efflux from cells.
Introduction High-affinity binding of high-density lipoprotein (HDL) by human [1,2] and rat [3,4] small intestine epithelial cells (enterocytes) has been recently demonstrated in our and other laboratories. This binding was accompanied with rapid degradation of HDL which made it possible to assume that small intestine may be the site of HDL catabolism. However, the experiments in vivo on rats failed to confirm the key role of the small Correspondence: D.D. Sviridov, Institute of Experimental Cardiology, USSR Cardiology Research Center, 3rd Cherepkovskaya Street 15, Moscow 121552, U.S.S.R.
0005-2760/87/$03.50
Q 1987 Elsevier Science Publishers
intestine in HDL catabolism [5,6]. Alternatively, it is possible to assume that enterocytes may use HDL as a source of cholesterol to synthesize chylomicrons when the diet is poor in cholesterol. In this case HDL interaction with enterocytes should be regulated by cell cholesterol content. In the previous paper we demonstrated that the binding of i2’I-labeled HDL, with enterocytes is of high affinity, specific, saturable, reversible and accompanied with an increase of cholesterol synthesis by these cells [l]. In the present paper we assess whether cholesterol in turn may regulate the lz51-labeled HDL, binding with enterocytes and the possible fate of the bound ‘251-labeled HDL,. We have demonstrated that the interaction
B.V. (Biomedical
Division)
261
of HDL with human enterocytes may be followed both by the degradation of HDL and cholesterol efflux from cells, and can be regulated by cholesterol loading to cells. Materials and Methods Cells
Small intestine (middle jejunum) of children aged from several days to 10 years was taken at autopsy within l-2 h after death. The predominant causes of death were congenital heart and brain defects; intestine from donors who had suffered from gut and blood diseases and metabolic disorders was not used. Enterocytes were isolated as previously described [l]. In brief, small intestine segments were everted over plastic rods, washed in minimum essential medium (Flow) and then in Dulbecco’s phosphate-buffered saline (Flow) without Ca2+ or Mg2+. The everted segments were incubated for 10 min at 37” C and then for 1 h at 25°C with shaking at 140 rpm in the medium consisting of 65.7% phosphatebuffered saline without Ca2+ or Mg2+ and 34.3% tridistilled water, and supplemented with 1% (w/v) of poly(viny1 pyrrolidone) (M, 40 000, Sigma). Released cells were sedimented and washed twice in minimum essential medium by centrifugation at 500 X g for 10 min and filtered through a nylon mesh. The cells yield was l-2 - lo7 cells per 10 cm of the intestine; the viability of cells was 80-90X according to the Trypan blue exclusion test. The detailed characterization of isolated enterocytes was reported previously [l]. Cell protein content was determined according to Bradford [7]. Cholesterol content of cells was determined according to Roscheau et al. [8] using a Boehringer Mannheim enzymatic kit. A total of 12 preparations from 12 donors was used for experiments. A comparison of the results obtained on enterocytes from different donors gave a between preparation coefficient of variation of about 20% for the binding and 50% for the degradation data, and the within sample coefficient of variation was about 10% for binding and degradation data. Liproproteins
Low-density lipoprotein (LDL, d = 1.019-1.050
g/cm3) and HDL, (d = 1.125-1.216 g/cm’) were isolated from plasma of healthy donors by sequential preparation ultracentrifugation at 105 000 X g [9]. The homogeneity of lipoproteins was checked by analytical ultracentrifugation. The protein composition of lipoproteins was characterized by electrophoresis of delipidated samples in a 11% SDS-polyacrylamide gel [lo]; no apolipoprotein E was found in the HDL, fraction. The concentration of lipoproteins was evaluated by the protein content which was measured according to Bradford [7]. The cholesterol content of lipoproteins was determined according to Roscheau et al. [8]. Iodination of HDL, was perfromed according to Bilheimer et al. [ll]. The specific radioactivity of the 12’1-labeled HDL, preparations was 200-500 cpm/ng protein, 99.4% of “‘1 being found in the trichloroacetic acid-precipitated fraction. Lipoprotein-deficient serum was obtained by repeated ultracentrifugation of fresh blood serum from healthy donors at the density over 1.25 g/ cm3 for48hat105OOOxg. ‘2sI-labeled HDL,
binding by enterocytes
The incubation mixture contained 0.5 . lo6 cells, 50 ~1 lipoprotein-deficient serum, 20 pg 1251labeled HDL, and minimum essential medium to a final volume of 0.5 ml. To determine non-specific binding, a lo-20-fold excess of unlabeled HDL, was added to the incubation mixture. In the experiments performed at 4”C, the mixture was incubated for 1.5 h in the wells of a ‘Multidish (Nunc) in a cold-box with shaking at 60 rpm in an orbital shaker. In the experiments carried out at 37”C, the mixture was placed in the wells and incubated in a CO, incubator (Forma Scientific) (5% CO2 95% air) for 3 h with shaking at 60 rpm. After incubation the cells were sedimented by centrifugation at 1100 X g for 5 min at 4’ C and washed three times with 10 ml of minimum essential medium containing 1 mg/ml bovine serum albumin (Sigma). To determine the amount of ‘251-labeled HDL, bound to cells at 4’ C, the final pellet was counted in a gamma counter (Compugamma, LKB). To measure the amount of 1251labeled HDL, bound and internalized at 37”C, the cells were washed one more time with Ca2+-, Mg 2+- free phosphate-buffered saline resuspended in 1 ml of 0.05% trypsin 0.002% EDTA (Flow) and
268
incubated for 5 min at 37°C. The trypsin action was arrested by addition of 100 ~1 fetal calf serum (Flow) and the cells were sedimented by centrifugation at 1100 X g for 5 min at 4O C. The radioactivity of the supematant (‘binding’) and pellet (‘internalization’) was determined in a gamma counter. The degradation was determined as a trichloroacetic acid-soluble, non-iodine radioactivity in the medium and measured as previously described [l]. To investigate the effect of cholesterol on binding, ~tem~tion and degradation of ‘251-labeled HDL,, the cells were preincubated for 1 h at 37’C in minimum essential medium supplemented with cholesterol (Sigma S-CH) dissolved in ethanol. The ~ncentration of ethanol in the medium did not exceed 1%; the corresponding amount of ethanol was added to the control samples. After preincubation, the cells were sedimented by centrifugation at 500 x g for 10 mm at 4 * C, washed twice with 10 ml of minimum essential medium and used for determination of binding, internalization and degradation of ‘2SI-labeled HDL,. Cholesterolefflux from enterocytes Cells were preincubated for 1 h at 3’7’ C in minimum essential medium containing 10% lipoprotein-deficient serum and 10 /.&i/ml [r4C]cholesterol (Amersham, specific radioactivity 50 mCi/mmol) or 100 pg/ml unlabeled cholesterol, both dissolved in ethanol. After preincubation, the cells were sedimented by centrifugation at 500 X g for 10 min at 4°C and washed four times with 10 ml minimum essential medium. The radioactivity found in the medium after the last wash did not exceed 1% of that associated with cells. After washing, the cells were resuspended in minimum essential medium containing 10% lipoprotein-deficient serum placed in the wells of a multidish and incubated at 37” C in a CO, incubator for the indicated periods of time in the presence or absence of HDL, (100 ag/ml~. Control incubations were carried out in a cell-free medium. When the incubation was over, the cells were sedimented by cent~ugation at 500 X g for 10 min at 4* C and resuspended in 0.5 ml distilled water. In the experiments with unlabeled cholesterol, its content both in the medium and cells was determined
according to Roscheau et al. [S]. In the experiments with [14C]cholesterol, aliquots both of the medium and cells were transferred into scintilation vials and counted on a beta counter (RackBeta, LKB). The rest of the medium samples and control HDL, preparation (3 mg of protein) were applied to a potassium bromide density gradient from 1.019 to 1.260 g/cm3 and centrifuged for 22 h at 15°C in a SW-40 rotor [12]. Fractions were collected by puncturing the bottoms of the tubes after which their density, absorbance at 280 nm and ~14C]cholesterol content were measured. Each experiment was done in duplicate and reproduced two or three times on the preparation of enterocytes obtained from the other donors. Representative experiments are reported.
Y
d
I
0
I
I
1
2
3
4
Preincubation time (h)
Fig. 1. The dependence of ‘2sI-labeIed HDL, binding by enterocytes on the time of preincubations with cholesterol. Cells were preincubated with a non-lipoprotein cholesterol (100 pg/ml) at 37 o C for the indicated periods of time. Excess of cholesterol was wasbed out and the cells were incubated with ‘2sI-labeled HDL, (40 &ml) for 1.5 h at 4°C. The amount of bound 1251-labeled HDL, was determined. The inset shows the time course of the [14C]dholesterol uptake by enterocytes. CMls were incubated with [t4C]cholesterol (100 rkg/m.l) at 37*C for the indicated periods of time. The excess of [‘4C]cholesterol was washed out and amount of the cell-associated 14C radioactivity was determined. Each point represents the mean data of duplicate determinations.
269
ResultS Preincubation of enterocytes with a non-lipoprotein cholesterol led to the stimulation of 1251labeled HDL, binding by cells. The dependence of the effect on the pre~cubation time is shown in Fig. 1. The enhancement was most pronounced within the first 2 h. The time-course of the enhancement of ‘251-labeled HDL, binding correlates well with that of the [‘4C]cholesterol uptake by enterocytes (r = 0.95, P < 0.001) (Fig. 1, inset). Non-specific binding did not change witbin at least 2 h of preincubation with a non-lipoprotein cholesterol (not shown). The dependence of 125I-labeled HDL, binding on cholesterol concentration in the preincubation medium is presented in Fig. 2A. The binding grew sharply as cholesterol concentration was raised to 60 ~g/ml, while further increase in cholesterol concentration had a less pronounced effect. The internalization slightly increased at cholesterol concentrations up to 60 ~g/ml, although further rise of cholesterol concentration did not result i.n
A
the stimulation of inte~~~ation (Fig. 2B). The dependence of ‘251-labeled HDL, degradation on cholesterol concentration had a double-phase character (Fig. 2C). At concentrations from 10 to 20 pg/ml, the degradation rate was rapidly elevated, but further increase in cholesterol concentration led to a decrease of the degradation rate. Three separate experiments gave the similar curves and significant changes (P c 0.001) in the binding and degradation. Preincubation of enterocytes with LDL brought about the same changes in ‘251-labeled HDL, binding, intem~a~on and degradation (not shown). The non-specific binding, internalization and degradation were not affected by cholesterol loading to cells and neither was degradation of ‘251-labeled HDL, in the conditioned media (not shown). ~ein~bation of enterocytes with a non-lipoprotein cholesterol led to an increase of the cellular cholesterol content up to 2.Sfold (Fig. 3A). ‘251-labeled HDL, binding correlates well with the cholesterol content of enterocytes (r = 0.95, P < 0.001) (Fig. 3B).
B
$2
1200-
200-
’
800* .
SO
100
160
50 Cholesterol
100 150 cont. (,uglml )
50
100
l!%
Fig. 2. The dependence of ‘251-labeled HDL, binding (A), internalization (B) and degradation (C) on the cholesterol concentration in the preincubation medium. Cells were preincubated with a non-lipoprotein cholesterol for 1 h at 37 * C. The excess of choIestero1 was washed out and the cells were incubated with ‘251-labeled HDL, (40 cg/ml) for 3 h at 37O C. The binding, internalization and degradation were determined. Each point represents the mean data of duplicate determinations.
IB
l
i// r =0.95
100
200
Cholesterol
300
400
0
cont. ( @i)
100
200
300
Cellular cholesterol content (@glmg cell protein)
Fig. 3. The dependence of the cellular cholesterol content on the cholesterol concentration in the medium {A) and its correlation with the enhancement of ‘*sI-labeled HDL, binding (B). Cells were incubated with non-lipoprotein cholesterol for 1 h at 37OC. The excess of cholesterol was washed out and the cellular cholesterol content was determined using an enzymatic assay. Each point represents the mean data of duplicate determinations. Data on panel B were calculated from that on panel A and Fig. 2A.
l .
/
0
/ ,- 100 1251-HDL, bound (ng/mg
cell protein)
0
i
--
0
1
--ci;
--
0
2
Preincubation Fig. 4. The dose-dependent specific ‘251-labeled HDL, binding given in Scatchard plots. Cells were preincubated for 1 h at 37 o C with (0) or without (0) non-lipoprotein cholesterol (100 ag/ml). The excess of cholesterol was washed out and the cells were incubated for 3 h at 37 o C with different concentration of ‘*%iabeIed HDL, in the presence or absence of a 20-fold excess of unlabeled HDL,. The specific binding was calculated as the total minus non-specific one measured in the presence of excess of unlabeled HDL,. Each point represents the mean data of duplicate determinations.
_..A
--
3
4
time (h)
Fig. 5. Effect of cycloheximide on the chol~terol-~du~d enhancement of ‘251-labeled HDL, binding. Cells were preincubated at 37’ C for the indicated periods of time with 100 ng/ml non-lipoprotein cholesterol (O), cholesterol plus 0.5 mM cycloheximide (0), and cycloheximide alone (A). The excess of reagents was washed out and the cells were incubated with t2sI-labeled HDL, (40 pg/ml) for 1.5 h at 4*C. The amount of bound ‘251-labeled HDL, was determined. Each point represents the mean data of duplicate determinations.
271
r
I
I
100
50
Time (mid Fig. 6. Effect of cycloheximide on [14C]cholesterol uptake by enterocytes. Cells were incubated at 37O C for the indicated periods of time with [‘4C]cholesterol (100 &ml) in the presence (0) or absence (0) of 0.5 mM cycloheximide. The excess of reagents was washed out and amount of cell-associated 14C radioactivity was determined. Each point represents the mean data of duplicate determinations.
Fig. 4 shows the Scatchard analysis of a concentration-dependent binding of ‘251-labeled HDL, by enterocytes after preincubation with or without non-lipoprotein cholesterol. F’reincubation of enterocytes with cholesterol (100 pg/ml) brought about the 2-fold increase of the number of binding sites (B,,= 1246 + 52 vs. 655 f 30 ng/mg cell protein, P -C0.001) and 1.7-fold increase of the binding affinity (Kd = 7.7 f 2.1 vs. 13.1 f 3.1 /Ag/ml, P < 0.05). Fig. 5 shows the effect of cholesterol on 1251labeled HDL, binding by enterocytes in the presence of cycloheximide, the inhibitor of protein synthesis in eucaryotic cells. Cells were preincubated with cycloheximide for up to 4 h at 37 o C. Since further incubation of cells for 3 h at 37” C resulted in a dramatic decrease of cell viability, the binding assay was performed for 1.5 h at 4’ C. It was shown previously that binding reached equilibrium after incubation for 1.5 h at 4°C [l]. It was shown that cycloheximide (0.5 mM) does inhibit the cholesterol effect on ‘251-labeled HDL, binding (Fig. 5). Cycloheximide alone had no effect on ‘251-labeled HDL, binding (Fig. 5). However, cycloheximide unexpectedly inhibited [14C]-
A
‘\
0
~--XL____-_-----0
Time(h) Fig. 7. The time course of HDL,-mediated efflux of cholesterol from enterocytes (A) and its appearance in the medium (B). Each point represent the mean data of duplicate determinations. 0, medium with HDL,; 0, medium without HDL,.
212
P2o 1.3
A260 0.3
1 a2
7
1.2
1 1.1
1.0 Number
of fraction
cholesterol to tissues and cells which utilize relatively large amounts of cholesterol, e.g., for steroid hormone synthesis [13-151, and to facilitate the removal from other cells and tissues of excess cholesterol [13,15,16]. The small intestine mucosa may use HDL for both types of process. Dietary lipids adsorbed in the small intestine enter the lymph within chylomicrons synthesized in enterocytes [17]. However, the cholesterol content of chylornicrons (both in esterified and non-esterified form) is relatively constant [17] and may significantly differ from that in the diet. This difference could be due to endogenous synthesis and exogenous sources of cholesterol or possible cholesterol efflux. These processes could be regulated in response to the cellular cholesterol content. It was demonstrated in this study that the binding of ‘251-labeled HDL, by human enterocytes is up-regulated in response to cholesterol loading to cells. Preincubation of cells with both non-lipoprotein and LDL cholesterol leads to stimulation of ‘251-labeled HDL, binding in a time- and dose-dependent manner (Figs. 1, 2A) and the enhancement of the number of binding sites (Fig. 4). Loading of cells with cholesterol also resulted in the changes of the binding constants; however, these changes were small and, in some experiments, even statistically insignificant. The stimulation of binding correlates with the cellular cholesterol content (Fig. 3) and is cycloheximide sensitive (Fig. 5). However, in our experiments, cycloheximide inhibited cholesterol uptake by enterocytes (Fig. 6). The same effect of cycloheximide on cholesterol uptake by fibroblasts was shown by Tabas and Tall [18]. The extent of inhibition of cholesterol uptake by cycloheximide is lower than inhibition of ‘251-labeled HDL, binding. However, these data show that cycloheximide may alter the cholesterol content of the cell membrane which in turn may affect the ‘251labeled HDL, binding. To what extent different effects of cycloheximide are responsible for inhibition of ‘251-labeled HDL, binding is difficult to evaluate. However, taking into consideration that the effect of cholesterol is too rapid to explain the 2-fold increase in the amount of binding sites by their de novo synthesis, it seems probable that preincubation of cells with cholesterol results in the exposure of pre-existing binding sites rather
LI >I
0
Fig. 8. Analysis of [‘4C]cholesterol released from enterocytes. The cells loaded with [14C]cholesterol were incubated for 3 h at 37OC in the presence or absence of HDL, (100 pg/ml). The cells were sedimented and medium was analysed by density gradient centrifugation. l, medium with HDL,; 0, medium without HDL,; n, density value; A, HDL,.
cholesterol uptake by enterocytes (Fig. 6). The difference between the effect of cholesterol on ‘251-labeled HDL, binding and degradation made it possible to assume that not all bound HDL, particles are necessarily internalized and degraded by enterocytes. Some HDL, may possibly dissociate into the medium and mediate cholesterol efflux from enterocytes. To test this possibility, the cells were loaded with unlabeled or [‘4C]-cholesterol and incubated in the presence or absence of HDL,. It was shown that HDL, stimulate a prompt efflux both of unlabeled (Fig. 7) and [‘4C]cholesterol (not shown) from enterocytes into the medium: about 60% of cholesterol was released within the first hour (Fig 7). The released cholesterol was incorporated mostly into the particles of HDL density (Fig. 8). In the other experiments, cells not loaded with cholesterol were incubated with or without HDL,. HDL, stimulated release of cholesterol to the medium in a similar manner; however, the amount of released cholesterol was 2-fold lower than after loading of cells with cholesterol (not shown). Discussion
In terms of the number of circulating lipoprotein particles. HDL is the predominant plasma lipoprotein class in humans [13]. It is assumed that the functional role of HDL is to supply
213
than their synthesis de novo. The regulation of cell
functions on the membrane level is a well-documented phenomenon [19]. A similar effect of cholesterol on HDL, binding was shown for fibroblasts and smooth muscle cells [20]. It has been shown in the recent work of Kagami et al. [21] that alterations in the cholesterol uptake in the intestine of rats in vivo does not change the binding of HDL, by isolated enterocytes. However, these alterations were not accompanied by changes in the cell cholesterol content which makes it difficult to compare these data with our results. The change in the degradation rate has a twophase character: it increases at low cholesterol concentrations and returns to the control level at high cholesterol concentrations (Fig. 2C). The degradation of ‘251-labeled HDL, by enterocytes seems to be an intercellular process rather than an action of extracellular proteinases, since (i) it is chloroquine sensitive [l], (ii) the nonspecific degradation is not affected by loading of cells with cholesterol and (iii) degradation of ‘251-labeled HDL, in the conditioned medium was negligible [l] and was not affected by loading of cells with cholesterol. It is possible that the increase in the degradation rate is related to an increase in the number of binding sites, that is the amount of HDL, which has entered the cells. It was shown previously that the rate of ‘251-labeled HDL, degradation possibly far exceeds that of their internalization by enterocytes [l]. Therefore, alterations in the internalization rate are not necessarily accompanied with the corresponding changes in the amount of internalized ‘251-labeled HDL, (Fig. 2B,C). A further increase in the cellular cholesterol content leads to a situation where the excess of bound HDL, is no longer degraded. This, in turn, may be related to changes in the properties of the cell membranes or to specific functioning of the regulatory mechanisms, i.e., excessive accumulation of cholesterol apparently make HDL, serve as its acceptor. The latter possibility was suggested by the stimulation of cholesterol release from enterocytes by HDL,, which was measured both by the isotopic technique and cholesterol mass analysis (Fig. 7). The HDL-mediated efflux of cholesterol from cells is a well-established phenomenon (for review see Refs. 13,15).
It may be explained by both uptake of
cholesterol from cell surface membrane and retroendocytosis. Our results do not permit us to distinguish between these possibilities; however, since the process is very rapid the first mechanism seems more probable. In conclusion, the cellular cholesterol content seems to be at least one of the factors which regulate HDL interactions with human enterocytes. Acknowledgements The authors acknowledge the assistance of V.O. Ivanov in iodination of lipoprotein, and the staff of the Pathology Department of the 1st Moscow Children’s Hospital in performing autopsies. References 1 Sviridov, D.D., Safonova, I.G., Gusev, V.A., TaIaIaev, A.G., Tsybulsky, V.P., Ivanov, V.O., Preobrazensky, S.N., Repin, VS. and Smimov, V.N. (1986) Metabolism 35, 588-595 2 Chazov, E.I., Sviridov, D.D., Safonova, LG., Ivanov, V.O. and Repin, V.S. (1984) Circulation 70,11-142 3 Suzuki, N., Fidge, N., Nestel, P. and Yin, J. (1983) J. Lipid. Res. 24253-264 4 Kagami, A., Fidge, N., Suzuki, N. and Nestel, P. (1984) B&him. Biophys. Acta 795, 179-190 5 Glass, K.S., Pittman, R.C., Keller, G.A. and Steinberg, D. (1983) J. Biol. Chem. 258, 7161-7167 6 Koelz, H.R., SheniB, B.C., Turley, S.D. and Dietschy, J.M. (1982) J. Biol. Chem. 257, 8061-8072 7 Bradford, M. (1976) Anal. B&hem. 72, 248-254 8 Roscheau, P., Bemt, E. and Gruber, W., (1974) Z. Khn. Chem. Khn. B&hem. 12,403-407 9 Lindgren, F.T. (1975) in Analysis of Lipids and Lipoproteins (Perkins, E.G., ed.), pp. 204-224, American Oil Chemists Society, New York 10 Stephens, R.E. (1975) Anal B&hem. 65, 369-379 11 BiIheimer, D.W., Eisenberg. S. and Levy, R. (1972) Biochim. Biophys. Acta 260, 212-221 12 Redgrave, T.G., Roberts, D.C.K. and West, C.E. (1975) Anal. B&hem. 65,42-49 13 Eisenberg, S. (1984) J. Lipid. Res. 25, 1017-1057 14 Gwynne, J.T. and Hess. B. (1980) J. Biol. Chem. 255, 10875-10883 15 Pittman, R.C. and Steinberg, D. (1984) J. Lipid. Res. 25, 1577-1585 16 Miller, N.E., LaViie, A. and Crook, D. (1985) Nature 314, 109-111 17 Green, P.H.R. and Ghckman, R.H. (1981) J. Lipid. Res. 22, 1153-1173 18 Tabas, I. and Tall, A.R., (1984) J. Biol. Chem. 259, 13897-13905
214
19 Spector, A.A. and Yorek, M.A. (1985) J. Lipid. Res. 26, 1015-1035 20 Oram, J.F., Brinton, E.A. and Bierman, E.L. (1983) J. Clin. Invest. 72, 1611-1621
21 Kagami, A., Fidge, Res. 26, 705-712
N.H.
and Nestel,
P.J. (1985) J. Lipid
Biochimicu
et Blophysica Am
919 (1981) 266-274 Elsevier
BBA 52540
Cholesterol regulates high-density lipoprotein interaction with isolated epithelial cells of human small intestine D.D. Sviridov a, I.G. Safonova a, V.P. Tsybulsky a, A.G. Talalaev ‘, S.N. Preobrazensky a, V.S. Repin a and V.N. Srnirnov a r?~nstif~te of ExFe~i~e~ta~ Cardioluw, and 6 Pathology Depariment, 1st Moscow
(Revised
Key words:
USSR
Card&lo@ Research Center
Chiidren‘s Hospital, ~Uoseow (U.S. S. R.)
(Received 12 June 1986) manuscript received 25 February
HDL binding;
Cholesterol;
Lipoprotein;
1987)
(Epithelial
cell)
The effect of choksterol on the interaction of ~gh-densi~ lipoprotein (HDL) with isolated haan smah intestine epitbelial cells (enterocytes) was studied. ‘251-labeled HDL, binding by these cells was enhanced in response to cholesterol loading of the cells in a time- and dose- dependent manner. Preincubation of the cells with cholesterol led to the enhancement both of the number of binding sites and the binding affinity. The enhancement of binding correiated with the cellular cholesterol content. Cycloheximide (0.5 mM) inhibited uptahe of cholesterol by enterocytes and’bloched its effect on ‘251-labeled HDL, binding. The effect of cholesterol on ‘251-labeled HDL, degradation had a double-phase character. At concentrations lo-20 pg/ml, the degradation rate was rapidly elevated, but further increase in cholesterol ~oneen~tion led to a fail in the degradation rate. Incubation of enteroqtes with HDL, resulted in the efflux of cholesterol from cells and its incorporation into HDL,. The results obtained mahe it possible to assume that binding and degradation of ‘251-labeled HDL, by human enterocytes are independently regulated by the eel1 total cholesteroI content. Binding of HDL by enterocytes may result both in the degradation of HDL and cholesterol efflux from cells.
Introduction High-affinity binding of high-density lipoprotein (HDL) by human [1,2] and rat [3,4] small intestine epithelial cells (enterocytes) has been recently demonstrated in our and other laboratories. This binding was accompanied with rapid degradation of HDL which made it possible to assume that small intestine may be the site of HDL catabolism. However, the experiments in vivo on rats failed to confirm the key role of the small Correspondence: D.D. Sviridov, Institute of Experimental Cardiology, USSR Cardiology Research Center, 3rd Cherepkovskaya Street 15, Moscow 121552, U.S.S.R.
0005-2760/87/$03.50
Q 1987 Elsevier Science Publishers
intestine in HDL catabolism [5,6]. Alternatively, it is possible to assume that enterocytes may use HDL as a source of cholesterol to synthesize chylomicrons when the diet is poor in cholesterol. In this case HDL interaction with enterocytes should be regulated by cell cholesterol content. In the previous paper we demonstrated that the binding of i2’I-labeled HDL, with enterocytes is of high affinity, specific, saturable, reversible and accompanied with an increase of cholesterol synthesis by these cells [l]. In the present paper we assess whether cholesterol in turn may regulate the lz51-labeled HDL, binding with enterocytes and the possible fate of the bound ‘251-labeled HDL,. We have demonstrated that the interaction
B.V. (Biomedical
Division)
261
of HDL with human enterocytes may be followed both by the degradation of HDL and cholesterol efflux from cells, and can be regulated by cholesterol loading to cells. Materials and Methods Cells
Small intestine (middle jejunum) of children aged from several days to 10 years was taken at autopsy within l-2 h after death. The predominant causes of death were congenital heart and brain defects; intestine from donors who had suffered from gut and blood diseases and metabolic disorders was not used. Enterocytes were isolated as previously described [l]. In brief, small intestine segments were everted over plastic rods, washed in minimum essential medium (Flow) and then in Dulbecco’s phosphate-buffered saline (Flow) without Ca2+ or Mg2+. The everted segments were incubated for 10 min at 37” C and then for 1 h at 25°C with shaking at 140 rpm in the medium consisting of 65.7% phosphatebuffered saline without Ca2+ or Mg2+ and 34.3% tridistilled water, and supplemented with 1% (w/v) of poly(viny1 pyrrolidone) (M, 40 000, Sigma). Released cells were sedimented and washed twice in minimum essential medium by centrifugation at 500 X g for 10 min and filtered through a nylon mesh. The cells yield was l-2 - lo7 cells per 10 cm of the intestine; the viability of cells was 80-90X according to the Trypan blue exclusion test. The detailed characterization of isolated enterocytes was reported previously [l]. Cell protein content was determined according to Bradford [7]. Cholesterol content of cells was determined according to Roscheau et al. [8] using a Boehringer Mannheim enzymatic kit. A total of 12 preparations from 12 donors was used for experiments. A comparison of the results obtained on enterocytes from different donors gave a between preparation coefficient of variation of about 20% for the binding and 50% for the degradation data, and the within sample coefficient of variation was about 10% for binding and degradation data. Liproproteins
Low-density lipoprotein (LDL, d = 1.019-1.050
g/cm3) and HDL, (d = 1.125-1.216 g/cm’) were isolated from plasma of healthy donors by sequential preparation ultracentrifugation at 105 000 X g [9]. The homogeneity of lipoproteins was checked by analytical ultracentrifugation. The protein composition of lipoproteins was characterized by electrophoresis of delipidated samples in a 11% SDS-polyacrylamide gel [lo]; no apolipoprotein E was found in the HDL, fraction. The concentration of lipoproteins was evaluated by the protein content which was measured according to Bradford [7]. The cholesterol content of lipoproteins was determined according to Roscheau et al. [8]. Iodination of HDL, was perfromed according to Bilheimer et al. [ll]. The specific radioactivity of the 12’1-labeled HDL, preparations was 200-500 cpm/ng protein, 99.4% of “‘1 being found in the trichloroacetic acid-precipitated fraction. Lipoprotein-deficient serum was obtained by repeated ultracentrifugation of fresh blood serum from healthy donors at the density over 1.25 g/ cm3 for48hat105OOOxg. ‘2sI-labeled HDL,
binding by enterocytes
The incubation mixture contained 0.5 . lo6 cells, 50 ~1 lipoprotein-deficient serum, 20 pg 1251labeled HDL, and minimum essential medium to a final volume of 0.5 ml. To determine non-specific binding, a lo-20-fold excess of unlabeled HDL, was added to the incubation mixture. In the experiments performed at 4”C, the mixture was incubated for 1.5 h in the wells of a ‘Multidish (Nunc) in a cold-box with shaking at 60 rpm in an orbital shaker. In the experiments carried out at 37”C, the mixture was placed in the wells and incubated in a CO, incubator (Forma Scientific) (5% CO2 95% air) for 3 h with shaking at 60 rpm. After incubation the cells were sedimented by centrifugation at 1100 X g for 5 min at 4’ C and washed three times with 10 ml of minimum essential medium containing 1 mg/ml bovine serum albumin (Sigma). To determine the amount of ‘251-labeled HDL, bound to cells at 4’ C, the final pellet was counted in a gamma counter (Compugamma, LKB). To measure the amount of 1251labeled HDL, bound and internalized at 37”C, the cells were washed one more time with Ca2+-, Mg 2+- free phosphate-buffered saline resuspended in 1 ml of 0.05% trypsin 0.002% EDTA (Flow) and
268
incubated for 5 min at 37°C. The trypsin action was arrested by addition of 100 ~1 fetal calf serum (Flow) and the cells were sedimented by centrifugation at 1100 X g for 5 min at 4O C. The radioactivity of the supematant (‘binding’) and pellet (‘internalization’) was determined in a gamma counter. The degradation was determined as a trichloroacetic acid-soluble, non-iodine radioactivity in the medium and measured as previously described [l]. To investigate the effect of cholesterol on binding, ~tem~tion and degradation of ‘251-labeled HDL,, the cells were preincubated for 1 h at 37’C in minimum essential medium supplemented with cholesterol (Sigma S-CH) dissolved in ethanol. The ~ncentration of ethanol in the medium did not exceed 1%; the corresponding amount of ethanol was added to the control samples. After preincubation, the cells were sedimented by centrifugation at 500 x g for 10 mm at 4 * C, washed twice with 10 ml of minimum essential medium and used for determination of binding, internalization and degradation of ‘2SI-labeled HDL,. Cholesterolefflux from enterocytes Cells were preincubated for 1 h at 3’7’ C in minimum essential medium containing 10% lipoprotein-deficient serum and 10 /.&i/ml [r4C]cholesterol (Amersham, specific radioactivity 50 mCi/mmol) or 100 pg/ml unlabeled cholesterol, both dissolved in ethanol. After preincubation, the cells were sedimented by centrifugation at 500 X g for 10 min at 4°C and washed four times with 10 ml minimum essential medium. The radioactivity found in the medium after the last wash did not exceed 1% of that associated with cells. After washing, the cells were resuspended in minimum essential medium containing 10% lipoprotein-deficient serum placed in the wells of a multidish and incubated at 37” C in a CO, incubator for the indicated periods of time in the presence or absence of HDL, (100 ag/ml~. Control incubations were carried out in a cell-free medium. When the incubation was over, the cells were sedimented by cent~ugation at 500 X g for 10 min at 4* C and resuspended in 0.5 ml distilled water. In the experiments with unlabeled cholesterol, its content both in the medium and cells was determined
according to Roscheau et al. [S]. In the experiments with [14C]cholesterol, aliquots both of the medium and cells were transferred into scintilation vials and counted on a beta counter (RackBeta, LKB). The rest of the medium samples and control HDL, preparation (3 mg of protein) were applied to a potassium bromide density gradient from 1.019 to 1.260 g/cm3 and centrifuged for 22 h at 15°C in a SW-40 rotor [12]. Fractions were collected by puncturing the bottoms of the tubes after which their density, absorbance at 280 nm and ~14C]cholesterol content were measured. Each experiment was done in duplicate and reproduced two or three times on the preparation of enterocytes obtained from the other donors. Representative experiments are reported.
Y
d
I
0
I
I
1
2
3
4
Preincubation time (h)
Fig. 1. The dependence of ‘2sI-labeIed HDL, binding by enterocytes on the time of preincubations with cholesterol. Cells were preincubated with a non-lipoprotein cholesterol (100 pg/ml) at 37 o C for the indicated periods of time. Excess of cholesterol was wasbed out and the cells were incubated with ‘2sI-labeled HDL, (40 &ml) for 1.5 h at 4°C. The amount of bound 1251-labeled HDL, was determined. The inset shows the time course of the [14C]dholesterol uptake by enterocytes. CMls were incubated with [t4C]cholesterol (100 rkg/m.l) at 37*C for the indicated periods of time. The excess of [‘4C]cholesterol was washed out and amount of the cell-associated 14C radioactivity was determined. Each point represents the mean data of duplicate determinations.
269
ResultS Preincubation of enterocytes with a non-lipoprotein cholesterol led to the stimulation of 1251labeled HDL, binding by cells. The dependence of the effect on the pre~cubation time is shown in Fig. 1. The enhancement was most pronounced within the first 2 h. The time-course of the enhancement of ‘251-labeled HDL, binding correlates well with that of the [‘4C]cholesterol uptake by enterocytes (r = 0.95, P < 0.001) (Fig. 1, inset). Non-specific binding did not change witbin at least 2 h of preincubation with a non-lipoprotein cholesterol (not shown). The dependence of 125I-labeled HDL, binding on cholesterol concentration in the preincubation medium is presented in Fig. 2A. The binding grew sharply as cholesterol concentration was raised to 60 ~g/ml, while further increase in cholesterol concentration had a less pronounced effect. The internalization slightly increased at cholesterol concentrations up to 60 ~g/ml, although further rise of cholesterol concentration did not result i.n
A
the stimulation of inte~~~ation (Fig. 2B). The dependence of ‘251-labeled HDL, degradation on cholesterol concentration had a double-phase character (Fig. 2C). At concentrations from 10 to 20 pg/ml, the degradation rate was rapidly elevated, but further increase in cholesterol concentration led to a decrease of the degradation rate. Three separate experiments gave the similar curves and significant changes (P c 0.001) in the binding and degradation. Preincubation of enterocytes with LDL brought about the same changes in ‘251-labeled HDL, binding, intem~a~on and degradation (not shown). The non-specific binding, internalization and degradation were not affected by cholesterol loading to cells and neither was degradation of ‘251-labeled HDL, in the conditioned media (not shown). ~ein~bation of enterocytes with a non-lipoprotein cholesterol led to an increase of the cellular cholesterol content up to 2.Sfold (Fig. 3A). ‘251-labeled HDL, binding correlates well with the cholesterol content of enterocytes (r = 0.95, P < 0.001) (Fig. 3B).
B
$2
1200-
200-
’
800* .
SO
100
160
50 Cholesterol
100 150 cont. (,uglml )
50
100
l!%
Fig. 2. The dependence of ‘251-labeled HDL, binding (A), internalization (B) and degradation (C) on the cholesterol concentration in the preincubation medium. Cells were preincubated with a non-lipoprotein cholesterol for 1 h at 37 * C. The excess of choIestero1 was washed out and the cells were incubated with ‘251-labeled HDL, (40 cg/ml) for 3 h at 37O C. The binding, internalization and degradation were determined. Each point represents the mean data of duplicate determinations.
IB
l
i// r =0.95
100
200
Cholesterol
300
400
0
cont. ( @i)
100
200
300
Cellular cholesterol content (@glmg cell protein)
Fig. 3. The dependence of the cellular cholesterol content on the cholesterol concentration in the medium {A) and its correlation with the enhancement of ‘*sI-labeled HDL, binding (B). Cells were incubated with non-lipoprotein cholesterol for 1 h at 37OC. The excess of cholesterol was washed out and the cellular cholesterol content was determined using an enzymatic assay. Each point represents the mean data of duplicate determinations. Data on panel B were calculated from that on panel A and Fig. 2A.
l .
/
0
/ ,- 100 1251-HDL, bound (ng/mg
cell protein)
0
i
--
0
1
--ci;
--
0
2
Preincubation Fig. 4. The dose-dependent specific ‘251-labeled HDL, binding given in Scatchard plots. Cells were preincubated for 1 h at 37 o C with (0) or without (0) non-lipoprotein cholesterol (100 ag/ml). The excess of cholesterol was washed out and the cells were incubated for 3 h at 37 o C with different concentration of ‘*%iabeIed HDL, in the presence or absence of a 20-fold excess of unlabeled HDL,. The specific binding was calculated as the total minus non-specific one measured in the presence of excess of unlabeled HDL,. Each point represents the mean data of duplicate determinations.
_..A
--
3
4
time (h)
Fig. 5. Effect of cycloheximide on the chol~terol-~du~d enhancement of ‘251-labeled HDL, binding. Cells were preincubated at 37’ C for the indicated periods of time with 100 ng/ml non-lipoprotein cholesterol (O), cholesterol plus 0.5 mM cycloheximide (0), and cycloheximide alone (A). The excess of reagents was washed out and the cells were incubated with t2sI-labeled HDL, (40 pg/ml) for 1.5 h at 4*C. The amount of bound ‘251-labeled HDL, was determined. Each point represents the mean data of duplicate determinations.
271
r
I
I
100
50
Time (mid Fig. 6. Effect of cycloheximide on [14C]cholesterol uptake by enterocytes. Cells were incubated at 37O C for the indicated periods of time with [‘4C]cholesterol (100 &ml) in the presence (0) or absence (0) of 0.5 mM cycloheximide. The excess of reagents was washed out and amount of cell-associated 14C radioactivity was determined. Each point represents the mean data of duplicate determinations.
Fig. 4 shows the Scatchard analysis of a concentration-dependent binding of ‘251-labeled HDL, by enterocytes after preincubation with or without non-lipoprotein cholesterol. F’reincubation of enterocytes with cholesterol (100 pg/ml) brought about the 2-fold increase of the number of binding sites (B,,= 1246 + 52 vs. 655 f 30 ng/mg cell protein, P -C0.001) and 1.7-fold increase of the binding affinity (Kd = 7.7 f 2.1 vs. 13.1 f 3.1 /Ag/ml, P < 0.05). Fig. 5 shows the effect of cholesterol on 1251labeled HDL, binding by enterocytes in the presence of cycloheximide, the inhibitor of protein synthesis in eucaryotic cells. Cells were preincubated with cycloheximide for up to 4 h at 37 o C. Since further incubation of cells for 3 h at 37” C resulted in a dramatic decrease of cell viability, the binding assay was performed for 1.5 h at 4’ C. It was shown previously that binding reached equilibrium after incubation for 1.5 h at 4°C [l]. It was shown that cycloheximide (0.5 mM) does inhibit the cholesterol effect on ‘251-labeled HDL, binding (Fig. 5). Cycloheximide alone had no effect on ‘251-labeled HDL, binding (Fig. 5). However, cycloheximide unexpectedly inhibited [14C]-
A
‘\
0
~--XL____-_-----0
Time(h) Fig. 7. The time course of HDL,-mediated efflux of cholesterol from enterocytes (A) and its appearance in the medium (B). Each point represent the mean data of duplicate determinations. 0, medium with HDL,; 0, medium without HDL,.
212
P2o 1.3
A260 0.3
1 a2
7
1.2
1 1.1
1.0 Number
of fraction
cholesterol to tissues and cells which utilize relatively large amounts of cholesterol, e.g., for steroid hormone synthesis [13-151, and to facilitate the removal from other cells and tissues of excess cholesterol [13,15,16]. The small intestine mucosa may use HDL for both types of process. Dietary lipids adsorbed in the small intestine enter the lymph within chylomicrons synthesized in enterocytes [17]. However, the cholesterol content of chylornicrons (both in esterified and non-esterified form) is relatively constant [17] and may significantly differ from that in the diet. This difference could be due to endogenous synthesis and exogenous sources of cholesterol or possible cholesterol efflux. These processes could be regulated in response to the cellular cholesterol content. It was demonstrated in this study that the binding of ‘251-labeled HDL, by human enterocytes is up-regulated in response to cholesterol loading to cells. Preincubation of cells with both non-lipoprotein and LDL cholesterol leads to stimulation of ‘251-labeled HDL, binding in a time- and dose-dependent manner (Figs. 1, 2A) and the enhancement of the number of binding sites (Fig. 4). Loading of cells with cholesterol also resulted in the changes of the binding constants; however, these changes were small and, in some experiments, even statistically insignificant. The stimulation of binding correlates with the cellular cholesterol content (Fig. 3) and is cycloheximide sensitive (Fig. 5). However, in our experiments, cycloheximide inhibited cholesterol uptake by enterocytes (Fig. 6). The same effect of cycloheximide on cholesterol uptake by fibroblasts was shown by Tabas and Tall [18]. The extent of inhibition of cholesterol uptake by cycloheximide is lower than inhibition of ‘251-labeled HDL, binding. However, these data show that cycloheximide may alter the cholesterol content of the cell membrane which in turn may affect the ‘251labeled HDL, binding. To what extent different effects of cycloheximide are responsible for inhibition of ‘251-labeled HDL, binding is difficult to evaluate. However, taking into consideration that the effect of cholesterol is too rapid to explain the 2-fold increase in the amount of binding sites by their de novo synthesis, it seems probable that preincubation of cells with cholesterol results in the exposure of pre-existing binding sites rather
LI >I
0
Fig. 8. Analysis of [‘4C]cholesterol released from enterocytes. The cells loaded with [14C]cholesterol were incubated for 3 h at 37OC in the presence or absence of HDL, (100 pg/ml). The cells were sedimented and medium was analysed by density gradient centrifugation. l, medium with HDL,; 0, medium without HDL,; n, density value; A, HDL,.
cholesterol uptake by enterocytes (Fig. 6). The difference between the effect of cholesterol on ‘251-labeled HDL, binding and degradation made it possible to assume that not all bound HDL, particles are necessarily internalized and degraded by enterocytes. Some HDL, may possibly dissociate into the medium and mediate cholesterol efflux from enterocytes. To test this possibility, the cells were loaded with unlabeled or [‘4C]-cholesterol and incubated in the presence or absence of HDL,. It was shown that HDL, stimulate a prompt efflux both of unlabeled (Fig. 7) and [‘4C]cholesterol (not shown) from enterocytes into the medium: about 60% of cholesterol was released within the first hour (Fig 7). The released cholesterol was incorporated mostly into the particles of HDL density (Fig. 8). In the other experiments, cells not loaded with cholesterol were incubated with or without HDL,. HDL, stimulated release of cholesterol to the medium in a similar manner; however, the amount of released cholesterol was 2-fold lower than after loading of cells with cholesterol (not shown). Discussion
In terms of the number of circulating lipoprotein particles. HDL is the predominant plasma lipoprotein class in humans [13]. It is assumed that the functional role of HDL is to supply
213
than their synthesis de novo. The regulation of cell
functions on the membrane level is a well-documented phenomenon [19]. A similar effect of cholesterol on HDL, binding was shown for fibroblasts and smooth muscle cells [20]. It has been shown in the recent work of Kagami et al. [21] that alterations in the cholesterol uptake in the intestine of rats in vivo does not change the binding of HDL, by isolated enterocytes. However, these alterations were not accompanied by changes in the cell cholesterol content which makes it difficult to compare these data with our results. The change in the degradation rate has a twophase character: it increases at low cholesterol concentrations and returns to the control level at high cholesterol concentrations (Fig. 2C). The degradation of ‘251-labeled HDL, by enterocytes seems to be an intercellular process rather than an action of extracellular proteinases, since (i) it is chloroquine sensitive [l], (ii) the nonspecific degradation is not affected by loading of cells with cholesterol and (iii) degradation of ‘251-labeled HDL, in the conditioned medium was negligible [l] and was not affected by loading of cells with cholesterol. It is possible that the increase in the degradation rate is related to an increase in the number of binding sites, that is the amount of HDL, which has entered the cells. It was shown previously that the rate of ‘251-labeled HDL, degradation possibly far exceeds that of their internalization by enterocytes [l]. Therefore, alterations in the internalization rate are not necessarily accompanied with the corresponding changes in the amount of internalized ‘251-labeled HDL, (Fig. 2B,C). A further increase in the cellular cholesterol content leads to a situation where the excess of bound HDL, is no longer degraded. This, in turn, may be related to changes in the properties of the cell membranes or to specific functioning of the regulatory mechanisms, i.e., excessive accumulation of cholesterol apparently make HDL, serve as its acceptor. The latter possibility was suggested by the stimulation of cholesterol release from enterocytes by HDL,, which was measured both by the isotopic technique and cholesterol mass analysis (Fig. 7). The HDL-mediated efflux of cholesterol from cells is a well-established phenomenon (for review see Refs. 13,15).
It may be explained by both uptake of
cholesterol from cell surface membrane and retroendocytosis. Our results do not permit us to distinguish between these possibilities; however, since the process is very rapid the first mechanism seems more probable. In conclusion, the cellular cholesterol content seems to be at least one of the factors which regulate HDL interactions with human enterocytes. Acknowledgements The authors acknowledge the assistance of V.O. Ivanov in iodination of lipoprotein, and the staff of the Pathology Department of the 1st Moscow Children’s Hospital in performing autopsies. References 1 Sviridov, D.D., Safonova, I.G., Gusev, V.A., TaIaIaev, A.G., Tsybulsky, V.P., Ivanov, V.O., Preobrazensky, S.N., Repin, VS. and Smimov, V.N. (1986) Metabolism 35, 588-595 2 Chazov, E.I., Sviridov, D.D., Safonova, LG., Ivanov, V.O. and Repin, V.S. (1984) Circulation 70,11-142 3 Suzuki, N., Fidge, N., Nestel, P. and Yin, J. (1983) J. Lipid. Res. 24253-264 4 Kagami, A., Fidge, N., Suzuki, N. and Nestel, P. (1984) B&him. Biophys. Acta 795, 179-190 5 Glass, K.S., Pittman, R.C., Keller, G.A. and Steinberg, D. (1983) J. Biol. Chem. 258, 7161-7167 6 Koelz, H.R., SheniB, B.C., Turley, S.D. and Dietschy, J.M. (1982) J. Biol. Chem. 257, 8061-8072 7 Bradford, M. (1976) Anal. B&hem. 72, 248-254 8 Roscheau, P., Bemt, E. and Gruber, W., (1974) Z. Khn. Chem. Khn. B&hem. 12,403-407 9 Lindgren, F.T. (1975) in Analysis of Lipids and Lipoproteins (Perkins, E.G., ed.), pp. 204-224, American Oil Chemists Society, New York 10 Stephens, R.E. (1975) Anal B&hem. 65, 369-379 11 BiIheimer, D.W., Eisenberg. S. and Levy, R. (1972) Biochim. Biophys. Acta 260, 212-221 12 Redgrave, T.G., Roberts, D.C.K. and West, C.E. (1975) Anal. B&hem. 65,42-49 13 Eisenberg, S. (1984) J. Lipid. Res. 25, 1017-1057 14 Gwynne, J.T. and Hess. B. (1980) J. Biol. Chem. 255, 10875-10883 15 Pittman, R.C. and Steinberg, D. (1984) J. Lipid. Res. 25, 1577-1585 16 Miller, N.E., LaViie, A. and Crook, D. (1985) Nature 314, 109-111 17 Green, P.H.R. and Ghckman, R.H. (1981) J. Lipid. Res. 22, 1153-1173 18 Tabas, I. and Tall, A.R., (1984) J. Biol. Chem. 259, 13897-13905
214
19 Spector, A.A. and Yorek, M.A. (1985) J. Lipid. Res. 26, 1015-1035 20 Oram, J.F., Brinton, E.A. and Bierman, E.L. (1983) J. Clin. Invest. 72, 1611-1621
21 Kagami, A., Fidge, Res. 26, 705-712
N.H.
and Nestel,
P.J. (1985) J. Lipid