Binding of high density lipoproteins to cell receptors promotes translocation of cholesterol from intracellular membranes to the cell surface.

Cultured cells have on their cell surface a specific high-affinity binding site (receptor) for high density lipoproteins (HDL) which appears to promote cholesterol efflux. In this study we characterized the cellular mechanisms involved in HDL receptor-mediated transport of cholesterol from cultured human fibroblasts and bovine aortic endothelial cells. HDL3, chemically modified by tetranitromethane (TNM-HDL3), is not recognized by this receptor and was used as a control for efflux not mediated by HDL receptor binding. HDL3 and TNM-HDL3 were found to be equally effective in causing efflux of plasma membrane cholesterol radiolabeled with [3H]cholesterol. However, HDL3 was much more effective than TNM-HDL3 in causing efflux of [3H]cholesterol associated with intracellular membranes. By measuring movement of endogenously synthesized [3H]cholesterol to the plasma membrane, and into the medium, we found that HDL3 induced a rapid movement of [3H]cholesterol from a preplasma membrane compartment to the plasma membrane that preceded [3H]cholesterol efflux. This effect was not observed with TNM-HDL3. Thus, receptor binding of HDL3 appears to facilitate removal of cellular cholesterol from specific intracellular pools by initiation of translocation of intracellular cholesterol to the plasma membrane.

from this laboratory have shown that several different extrahepatic cell types have receptors on their surface that bind HDL with high affinity and specificity (10)(11)(12). This receptor binding of HDL may facilitate the movement of free cholesterol from cells to HDL particles. The purpose of this study was to characterize the cellular mechanism involved in HDL receptor-mediated transport of cholesterol from cells.

EXPERIMENTAL PROCEDURES
Cells-Bovine aortic endothelial cells were obtained and cultivated as described by Schwartz (13). Human skin fibroblasts were grown as previously described (10, ll), using 10% fetal calf serum in Dulbecco's minimum essential medium (GIBCO). Cells were seeded in 35-mm dishes at a density of approximately 150,000 (endothelial cells) or 70,000 (fibroblasts) cells/dish. After 5-6 days, when confluent, the fibroblasts were washed once with pbosphate-buffered saline (PBS, pH 7.4) and were then incubated for 48 h with serumfree minimum essential medium (SF-MEM) containing 2 mg/ml fatty acid free albumin (FAFA) and 50 ,ug/ml cholesterol (dissolved in ethanol) to upregulate HDL, binding (11). The endothelial cells were incubated similarly, except that the incubation medium contained no FAFA and 30 pg/ml acetylated LDL instead of cholesterol (12).
Lipoproteins-Low density (LDL, d = 1.019-1.063 g/ml) and high density (HDLB, d = 1.125-1.21 g/ml) lipoproteins were isolated by standard sequential ultracentrifugation technique. HDL3 was iodinated with by the McFarlane monochloride procedure as modified for lipoproteins (15) to a specific activity of approximately 100-250 cpm/ng of protein. Unlabeled HDL3 was subjected to heparin-agarose affinity chromatography to remove apoB and apoE (10).
Modificatwn of Lipoproteins-LDL waa acetylated according to the method of Fraenkel-Conrat (16) as modified by Basu et al. (17). HDL3 was treated with tetranitromethane (TNM) as described previously (14). Briefly, 20 pl of a freshly prepared solution of 0.6 M TNM in absolute ethanol was added to 2 ml of HDL3 (2 mg/ml in 0.9% NaC1; final TNM concentration 6 mM) and the reaction was allowed to proceed in the dark at room temperature for 60 min. The sample was chilled on ice and was subjected to extensive dialysis against 0.9% NaCl with 1 mM EDTA. The resulting TNM-HDL, was passed through a 0.22-pm sterile filter and was kept at 4 'C in the dark.
HDL Binding-Binding of HDL3 to cultured fibroblasts was determined at 4 "C as described previously (11). Prior to the binding experiments, cells were loaded with non-lipoprotein cholesterol for 48 h as described earlier. The cells were chilled on ice and washed five times with cold PBS. Then 1.0 ml of cold SF-MEM with 10 mM HEPES, 1 mg/ml FAFA, and 2 pg/ml '%I-HDL3 was added to each dish. For competition studies, binding of "'I-HDL3 to cells was determined in the presence of a 20-fold excess of either HDL3 or TNM-HDL3. After 2 h at 0 "C the cells were washed three times with cold PBS containing albumin and two times with PBS. Cells were then digested in 0.1 M NaOH and an aliquot was assayed for "I radioactivity and another aliquot was assayed for protein content.  (18). ['HIMevalonolactone was used as a precursor of cholesterol because it bypasses the hydroxymethylglutaryl-CoA reductase step in cholesterol biosynthesis, the activity of which is suppressed in cholesterol-loaded cells. Cholesterol-loaded cells were chilled and washed five times with 2 ml of cold PBS. After the final wash, 1.0 ml of SF-MEM with 10 mM HEPES, 1 mg/ml FAFA, and 0.4 mM [3H]mevalonolactone (10 pCi/ ml, Du Pont-New England Nuclear) was added to each dish. The cells were then incubated in a 15 f 1 "C water bath for 6 h. After this pulse, the cells were chilled, washed five times on ice with cold PBS, and used for measurement of [3H]cholesterol efflux.
Labeling of the Plasma Membrane with PHlCholesterol-To enrich the plasma membrane cholesterol pool with endogenously synthesized [3H]cholesterol, cells were incubated with medium containing [,HI mevalonolactone as described above except that the incubations were done for 25 h at 37 "C. To enrich plasma membrane with [,HI cholesterol from an exogenous source, cholesterol-loaded cells were chilled on ice, washed five times with 2 ml of cold PBS, and then incubated for 6 h at 15 "C in SF-MEM containing 10 mM HEPES, 1 mg/ml FAFA, 0.4 mM mevalonolactone, and 0.2 pCi/ml [3H] cholesterol (55 Ci/mmol, Amereham Corp.). With this procedure, ['HI cholesterol is specifically incorporated into the plasma membranes of cells (19).
Measurement of pH/Cholesterol Effw-After pretreatment with either [3H]mevalonolactone or [3H]cholesterol, cells were chilled on ice and washed five times with cold PBS, and each dish received 1.0 ml of ice-cold SF-MEM with 10 mM HEPES, 1 mg/ml FAFA, 0.4 mM mevalonolactone, and indicated amounts of HDL, or TNM-HDL,. The dishes were then rapidly transferred to a 37 "C water bath and, after the time indicated, efflux media were collected. Each dish was then rinsed once with 1.0 ml of PBSFAFA, which was combined with the respective efflux medium. The cells were then chilled, washed 5 times with cold PBS (on ice), and fixed as described below. The tubes with the efflux media were centrifuged at 800 X g for 20 min (5 "C) to sediment detached cells, and the media were transferred to conical glass-stoppered tubes for extraction of [3H]cholesterol.
Treatment of CeUs with Cholesterol Oxidase-To measure the amount of endogenously synthesized (SH]cholesterol that had moved to the cell surface (plasma membrane) during an efflux incubation, a modification of the method reported by Lange and Ramos (19) for measurement of plasma membrane cholesterol was used.
In this procedure, after incubation and collection of efflux media, the cells were kept on ice and were washed an additional three times with cold PBS. Then 1 ml of cold 1% glutaraldehyde was added to each dish to fi the cells (10 min at 0 "C). After this step, the cells were washed another five times with cold PBS. Dishes were then transferred to a 37 ' C water bath, and 1.0 ml of PBS (37 "C) was added to each dish. After 2 min, 4 units of cholesterol oxidase (catalogue number 228250, Behring Diagnostics) was added (in 100 pl of PBS), and the cells were incubated with the enzyme for exactly 15 min. The cells were then rapidly chilled on ice and washed twice with cold PBS. The washed and fixed cell monolayers were immediately frozen and kept at -20 "C until analysis. Lipid Elctraction-[3H]Cholesterol and [3H]cholestenone were extracted from glutaraldehyde-fixed cells by hexane:isopropyl alcohol (3:2, v/v) as described (20).
[SH]Cholesterol was extracted from the efflux media according to the procedure of Folch et a l . (21). Sterol species were separated from each other on Silica Gel G thin-layer chromatography plates developed in heptane:diethyl etheamethanol (987.5:15, v/v/v) and detected with Iz. The individual spots were scraped into scintillation vials and were counted in Aquasol 2 (Du Pont-New England Nuclear).

Previous reports from
this laboratory have shown that tetranitromethane-modified HDL, particles (TNM-HDL,) lose their ability to compete with '%I-HDh for binding to the HDL receptor on fibroblasts and endothelial cells (14). Therefore, TNM-HDL3 was used in this study as a negative control for HDL3 to assess the effects of the receptor/HDL3 interaction on cellular cholesterol efflux. Before using the TNM-HDL, preparations, we routinely examined the ability of these modified HDL, particles to compete with 1261-HDL, for receptor binding. Typically, TNM-HDL, at a 20-fold excess displaced less than 15% of '=I-HDL, from the high-affinity binding sites, whereas the same excess of HDL, displaced more than 80% of the bound lZ5I-HDL3 (data not shown, cf.

Ref. 14).
To examine the ability of HDL, and TNM-HDL3 to cause efflux of plasma membrane cholesterol radiolabeled from an exogenous source, cells were pulsed with [3H]cholesterol (added in ethanol) and efflux was measured during a 4-h incubation a t 37 "C. Over the 4-h efflux period, both HDL, and TNM-HDL3 at concentrations between 5 and 100 Kg of protein/ml were equally effective in removing [3H]cholesterol from cells (Fig. l), indicating that receptor binding of HDL was not necessary for efflux of exogenously labeled plasma membrane cholesterol. Although efflux of plasma membranederived [3H]cholesterol appeared to be saturable with respect to the acceptor concentration, the apparent saturability may have been due to tracer depletion from the membrane pool rather than saturation of membrane binding sites, since more than 50% of the total cellular [3H]cholesterol was removed from cells during the 4-h experiment.
To test whether the interaction between HDL, and the cellsurface receptor leads to an enhanced efflux of newly synthe-  Fig. 2A). At 100 pg/ml of HDL3, more than 20% (total minus control) of the intracellular 3H label had been removed from the cells. At the same concentrations, TNM-HDL, was much less effective than HDL, in promotion of cholesterol efflux. When efflux of endogenously synthesized [3H]cholesterol was measured in endothelial cells using the same protocol, a similar efflux pattern could be observed ( mediated by TNM-HDL, appeared to be characterized mainly by a low-affinity component, since little saturability could be observed (Fig. 2, A and B ) . cholesterol in the plasma membrane and in the incubation medium showed that [3H]cholesterol first appeared on the cell surface but soon thereafter was found in the medium (Fig.   3, C and D, for fibroblasts and endothelial cells, respectively).
These results indicate that HDL3, but not TNM-HDL3, promotes movement of cholesterol from intracellular membranes to the cell surface, where it is subsequently released into the medium.
Receptor binding of HDL3 appears to specifically promote efflux of endogenously labeled cholesterol that resides within intracellular membranes. When incubations with 13H]mevalonolactone were extended to 25 h at 37 "C so that most of the [3H]cholesterol label was in the plasma membrane (as determined by cholesterol oxidase), HDL3 promoted [3H] cholesterol efflux by a nonsaturable, low-affinity process that was identical to that observed with TNM-HDL3 (Fig. 4). This finding is consistent with the efflux pattern shown in Fig. 1, where the plasma membrane cholesterol pool was labeled exogenously with [3H]cholesterol. Thus, removal of cholesterol from plasma membranes by HDL does not appear to be receptor-mediated, regardless of whether the plasma membrane cholesterol is derived from exogenous or endogenous sources.

DISCUSSION
Results in the present study show that the interaction between HDL, and its cell-surface receptor on cultured extrahepatic cells promotes selective movement of intracellular membrane-associated [3H]cholesterol to the plasma membrane and subsequent efflux of the translocated cholesterol. This succession of events was not seen in cells incubated with TNM-modified HDL,, which does not bind to the HDL receptor (14).
The observed HDL3-mediated [3H]cholesterol efflux appeared to be characterized by two different processes. First, as a result of receptor/HDL3 interaction, [3H]cholesterol was rapidly translocated from a preplasma membrane compartment to the plasma membrane. The second process was the actual efflux or removal of [3H]cholesterol from cells to the acceptor particles. Results suggest that the receptor/HDL, interaction was necessary only to trigger the initial translocation process. Once the [3H]cholesterol was in the plasma membrane, both HDL, and TNM-HDL, appeared to be equally effective in removing [3H]cholesterol from cells. This conclusion was based on studies showing that efflux of [,HI cholesterol already in the plasma membrane cholesterol pool was identical in the presence of both forms of HDL,. The lack of involvement of receptor binding in removal of cholesterol from plasma membranes is not surprising, since this process is thought to be mediated by spontaneous surface transfer (22). The endogenous [3H]cholesterol that was translocated and removed from cells following receptor binding of HDL, appears to be derived from the putative intracellular regulatory cholesterol pool, since fibroblasts incubated with HDL, show a significant and rapid down-regulation of the activity of the intracellular cholesterol esterifying enzyme, acyl-CoAcholesterol acyltransferase activity (5, 14). TNM-HDL, which neither binds to the HDL receptor nor stimulates cholesterol translocation, also does not down-regulate the activity of acyl-CoAcholesterol acyltransferase (14).
In the absence of HDL, the movement of endogenously synthesized [3H]cholesterol to the cellular plasma membrane was a slow process in cholesterol-loaded cells. This is in contrast to results obtained for cholesterol movement in cholesterol-depleted cells (23). It is possible that, with cholesterol-loaded cells, endogenously synthesized cholesterol is readily equilibrated with intracellular pools of excess free cholesterol, which are only partially translocated to the cell surface at a slow rate. Although HDLs accelerated movement of intracellular [3H]cholesterol to the cell surface in both fibroblasts and endothelial cells, the amount of [SH]cholesterol that was translocated and the kinetics of the processes were different in these two cell types, with endothelial cells translocating more cholesterol at a faster rate than fibroblasts. Despite these differences in rates of translocation, the rate of [3H]cholesterol efflux was similar in both cell types, suggesting that the efflux process per se may be rate-limiting.
Results in the present study do not allow conclusions to be drawn as to the mechanism of the HDL3-induced translocation of intracellular cholesterol to the cell surface. It has been proposed that, at least in macrophages, receptor binding of HDL can lead to endocytosis and resurfacing of receptor-HDL complexes, thus providing a targeting mechanism whereby HDL can remove cholesterol from specific intracel-lular compartments (24). However, this mechanism is an unlikely explanation for the effect of HDL, on cholesterol translocation since it has been shown that, under incubation conditions similar to those used in this study, internalization of HDL, could not be detected (25).
Thus, the results of this study suggest the existence of a specific biological response to the interaction between HDL, and its receptor. It is known that HDL, receptor activity is a function of total cellular unesterified cholesterol mass (11,12). These results suggest that the function of binding of HDL, to its receptor is to specifically promote transport of excess intracellular cholesterol to the cell surface where it can be removed from cells by HDL or other acceptor particles.