Efflux of cholesterol by high density lipoproteins (HDL) ()from extrahepatic tissue is thought to account for the protective effect of elevated HDL levels against atherosclerosis(1) . However, mechanisms by which HDL remove excess cholesterol from cells remain poorly understood, especially transport pathways involving removal of intracellular cholesterol. Removal of unesterified cholesterol from the plasma membrane has been adequately explained by the aqueous diffusion model. By this mechanism, cholesterol present in the plasma membrane desorbs from the cell surface to the surrounding aqueous space, diffusion occurs against a concentration gradient and the sterol molecule is absorbed by an appropriate extracellular acceptor, such as HDL(2, 3, 4) . The rate of desorption from cell membranes appears to be an intrinsic property of the membrane but also may be influenced by the cholesterol acceptor(3, 5, 6) . Kinetic analysis of cholesterol efflux data demonstrated that membrane cholesterol available for efflux exists in up to three kinetic pools with fast, intermediate, and slow desorption rates, depending on the cell type studied (reviewed in (7) ). These pools were proposed to arise from cholesterol distribution within different microdomains of the plasma membrane. Efflux of cholesterol from these pools appears to be a passive process although rates of cholesterol desorption from cell membranes could be modulated by apolipoproteins, primarily A-I(5, 8) , by altering distribution of cholesterol between the postulated domains. However, similar pools were shown to exist when phospholipid vesicles were used as the acceptors, suggesting that desorption of cholesterol from these domains is primarily a function of the membrane(7) . Such models for cholesterol efflux do not account for removal of cholesterol from other cellular compartments such as the pool(s) of sterol involved in the cholesteryl ester cycle or in the regulation of 3-hydroxy-3-methylglutaryl coenzyme A reductase and low density lipoprotein (LDL)-receptor activity(9) . Indeed, the cellular identities of such pools are poorly defined. Intracellular pools of cholesterol involved in the regulation of cholesterol homeostasis and the relationship between these pools and the various membrane pools have not been established.

Recent studies have examined the identity of the acyl coenzyme A:cholesterol acyl transferase (ACAT) substrate pool in cultured cells and have suggested that sterol available for esterification is derived from, or is in equilibrium with, plasma membrane cholesterol(10, 11) . These data imply transport of sterol from plasma membrane to sites of ACAT-mediated esterification in the endoplasmic reticulum must occur. Such results are consistent with data of Johnson et al.(12) , demonstrating that the majority of LDL-derived cholesterol is not directly esterified but instead rapidly transported to the plasma membrane from lysosomes, and little of the lysosomal cholesterol passes through the endoplasmic reticulum during transport to the cell membrane. Certain intracellular pools of cholesterol (i.e. not plasma membrane associated) can be depleted by HDL through a process that requires intact HDL apolipoproteins but are not affected by apolipoprotein-depleted HDL or artificial (non-apolipoprotein containing) acceptors(13, 14, 15, 16) . Additionally, several studies have shown that transport and efflux of cholesterol from intracellular sites can be activated by various signaling pathways, including activation of protein kinase C (17, 18) or protein kinase A (19, 20) under conditions that do not influence plasma membrane cholesterol desorption rates. These data suggest a dissociation between efflux of cholesterol present in the plasma membrane and intracellular pools based on sensitivity to different acceptor types or changes in cell metabolism.

These results suggest a transport pathway exists to deliver cholesterol from the plasma membrane to intracellular sites of cholesterol esterification and accumulation. Conversely, a pathway should also exist for the transport of free cholesterol (derived from cholesteryl ester hydrolysis and other sites of overaccumulation) within intracellular pools to the plasma membrane. Such a pathway has been shown to operate for HDL-mediated efflux of newly synthesized cholesterol in cholesterol overloaded cells(13, 18) . Also in support of this concept, a vesicular transport pathway has been demonstrated to exist for newly synthesized sterol molecules from the endoplasmic reticulum to the plasma membrane in growing cells(21) . Whether a similar mechanism(s) transports newly synthesized and excess cholesterol stored intracellularly to the plasma membrane for eventual efflux is unknown.

To better understand mechanisms of cholesterol transport between cellular compartments, the effects of various agents, known to affect cell protein transport pathways, were examined for their influence on cell cholesterol homeostasis and HDL-mediated cholesterol efflux. Results showed that monensin and brefeldin A caused a redistribution of cell cholesterol from the plasma membrane to intracellular compartments, and inhibited efflux of cholesterol available for removal by HDL. These compounds are known to alter the structure and function of the Golgi apparatus, and thus may implicate this intracellular organelle in cholesterol transport pathways influenced by HDL.

EXPERIMENTAL PROCEDURES

Materials

Methods

Cells

To label cell cholesterol pools to constant specific activity, subconfluent cultures were maintained in DMEM containing 10% FBS and 0.2 µCi/ml [H]cholesterol until confluent (usually 3 days). Labeled cells were subsequently loaded with non-lipoprotein cholesterol by incubation with DMEM containing 2 mg/ml fatty acid free bovine serum albumin (BSA) and 30 µg/ml cholesterol (added from an ethanol stock solution) for 24 h. Cultures were incubated for an additional 48 h in DMEM containing 1 mg/ml BSA, after which free and esterified pools of cholesterol attained constant specific activity (data not shown). Cells were also enriched with cholesterol by incubation with LDL. Subconfluent cultures were incubated with DMEM containing 2% lipoprotein-deficient FBS (LPDS, prepared by ultracentrifugation of FBS at density 1.25 g/ml and adjusted to equal protein content as FBS) containing 100 µg/ml LDL protein. After 48 h, by which time cells reached confluence, cells were rinsed with phosphate buffered saline (PBS) and incubated for an additional 24 h in DMEM containing 1 mg/ml BSA to allow equilibration of cholesterol pools. During incubation of cells with test compounds, vehicle alone (ethanol) was added to control dishes at equal concentrations (never exceeding 0.25%).

Lipoproteins

Cholesterol Esterification

[H]Cholesterol Efflux

Cholesterol Oxidase Treatment of Cells

Sterol Synthesis from [C]Acetate

Cholesterol Mass Efflux

Statistics

Other Methods

RESULTS

Cholesterol Esterification

The dose response of monensin and brefeldin A on cholesterol esterification was examined (Fig. 1). Increasing concentrations of monensin or brefeldin A had no effect on the basal rate of cholesterol esterification compared with controls during the 6-h incubation. In contrast, when HDL was present, monensin and brefeldin A inhibited the ability of HDL to decrease cellular cholesterol esterification. The effects of monensin were apparent at the lowest concentration tested and increased with dose, completely blocking the decrease in cholesterol esterification by HDL at 50 µM. Brefeldin A also prevented the decrease of cholesterol esterification by HDL at the lowest dose examined (0.2 µM) without any further effect up to 36 µM, inhibiting up to 70% of the decrease in cholesterol esterification by HDL obtained in the absence of brefeldin A. The dose response of HDL on cell cholesterol esterification in the presence or absence of these compounds was examined (Fig. 2). Under control conditions, HDL decreased ACAT activity in a concentration dependent manner. When monensin or brefeldin A was present, the decrease in cholesterol esterification by HDL was prevented at nearly all concentrations. Cholesterol esterification was inhibited an average of 56 ± 13% and 63 ± 9% for monensin and brefeldin A, respectively, over the tested concentration range. Removal of these compounds by washing the cells restored the ability of HDL to deplete ACAT substrate to control levels, demonstrating the effects of these compounds were reversible (data not shown). In these and other experiments, neither monensin or brefeldin A showed any cytotoxic effects compared with control cells based on recovery of cell protein, lactate dehydrogenase release to the medium, trypan blue dye exclusion, or incorporation of [C]oleate into phospholipids (not shown). Monensin, nigericin, and brefeldin A had similar effects on HDL-mediated ACAT inhibition when tested in porcine smooth muscle cells (data not shown), with the notable exception that monensin and nigericin decreased basal esterification rates by about 25% compared with controls.

Figure 1: Dose response of monensin and brefeldin A on cholesterol esterification in cholesterol-loaded fibroblasts in the presence or absence of HDL. Human skin fibroblasts were grown to confluence then loaded with non-lipoprotein cholesterol as described under ``Methods.'' Cultures were then incubated in serum-free DMEM containing 1 mg/ml BSA (, SFM) alone or with 50 µg/ml HDL protein (, +HDL) and the indicated concentrations of monensin or brefeldin A for 6 h at 37 °C. Cells were rinsed with PBS and incubated in DMEM containing 9 µM [C]oleate and 3 µM BSA for 1 h to measure cholesterol esterification as described under ``Methods.'' Results are expressed as picomoles of [C]oleate incorporated into [C]cholesterol esters/mg of cell protein. In A results are the mean ± S.D. of three dishes, and in B results are the mean of duplicate dishes, and the coefficient of variation was 7.8%.

Figure 2: Effect of monensin and brefeldin A on the decrease in cholesterol esterification by HDL in cholesterol-loaded fibroblasts. Human skin fibroblasts were grown and enriched with cholesterol as described in the legend to Fig. 1. Cultures were then incubated in serum-free DMEM containing 1 mg/ml BSA alone, 25 µM monensin, or 3.6 µM brefeldin A and the indicated concentrations of HDL for 16 h at 37 °C. Cholesterol esterification was subsequently measured as described in the legend to Fig. 1and expressed as picomoles of [C]oleate incorporated into [C]cholesterol esters/mg of cell protein. A: , control; , 25 µM monensin. Results are the mean of duplicate dishes, and the coefficient of variation was 5.6%. B: , control; , 3.6 µM brefeldin A. Results are the mean ± S.D. of three dishes, and missing error bars are contained within the symbols.

The time that monensin addition would inhibit the HDL-mediated decrease in ACAT activity was studied (Fig. 3). Addition of monensin at the start of the incubation or after 1 h maximally inhibited the decrease in cholesterol esterification due to HDL (70% of control values). Addition of monensin after 2 or 3 h showed partial inhibition (46 and 28%, respectively), and if monensin was added during the last hour of incubation with HDL, no effects were observed. Monensin had no statistically significant effect on the basal rate of cholesterol esterification in cultures incubated without HDL. These results show that monensin can effectively block HDL-mediated ACAT inhibition even after stimulation of this process had already begun.

Figure 3: Effect of addition time on the ability of monensin to prevent HDL-mediated decrease in cholesterol esterification in cholesterol-loaded human skin fibroblasts. Fibroblast cultures were grown and enriched with cholesterol as described in the legend to Fig. 1. Cultures were then incubated in serum-free DMEM containing 1 mg/ml BSA (, SFM) alone or containing 50 µg/ml HDL protein (, HDL) for a total of 6 h at 37 °C. Monensin was added to a final concentration of 25 µM at the indicated times. After incubation cells were incubated in DMEM containing [C]oleate as described in the legend to Fig. 1. Results are expressed as picomoles of [C]oleate incorporated into [C]cholesterol esters/mg of cell protein and are the mean ± S.D. of three dishes, missing error bars are contained within the symbols. The asterisk indicates p < 0.05 compared with control incubations containing HDL.

Cholesterol esterification in cells depleted of cholesterol and the ability of LDL to increase cholesterol esterification rates were examined after incubating cells with monensin, nigericin, or brefeldin A (Table 2). In cholesterol-depleted cells, HDL had a limited capacity to decrease cholesterol esterification (<10% compared with control medium) and was not examined. When ACAT substrate was limited, monensin and nigericin decreased cholesterol esterification by about 50%, suggesting that treatment of cells with these drugs diverts cholesterol out of the pool available for esterification. Brefeldin A increased esterification by 1.4-fold, possibly by increasing the availability of cholesterol to ACAT. Inclusion of LDL during the incubation caused a 6-fold increase in cholesterol esterification. As expected, monensin and nigericin blocked the uptake of LDL derived cholesterol(31) . Brefeldin A-treated cells processed LDL-derived cholesterol to a similar extent as control cells, suggesting that brefeldin A did not alter the uptake and lysosomal processing of LDL-cholesterol to sites of ACAT-mediated esterification under these conditions.

Stimulation of Sterol Synthesis by HDL

Figure 4: Effects of monensin and brefeldin A on stimulation of sterol synthesis by HDL in cholesterol-loaded human skin fibroblasts. Preconfluent fibroblast cultures were incubated for 48 h with DMEM containing 2% LPDS and 100 µg/ml LDL protein to enrich cholesterol pools followed by incubation for 24 h in DMEM containing 1 mg/ml BSA to allow equilibration of cholesterol pools. Cultures were incubated with DMEM containing 1 mg/ml BSA and the indicated concentrations of HDL protein alone (, Control), with 4 µM brefeldin A () or 25 µM monensin () for 24 h at 37 °C. Cells were then rinsed with PBS and incubated with DMEM containing 2 µCi/ml [C]acetate for 2 h at 37 °C. Cell lipids were extracted and saponified as described under ``Methods,'' and incorporation of radioactivity into sterols was measured after TLC separation of the non-saponified lipid fraction. Results are expressed as picomoles of [C]acetate incorporated into [C]sterols/mg of cell protein and are the mean ± S.D. of three dishes. Missing error bars are contained within the symbols.

Cell [H]Cholesterol Efflux and Sensitivity to Cholesterol Oxidase

Figure 5: Effects of monensin and brefeldin A on HDL-mediated cholesterol efflux and cell cholesterol distribution examined with cholesterol oxidase in cholesterol-loaded human skin fibroblasts. Fibroblast cultures were labeled with [H]cholesterol then loaded with non-lipoprotein cholesterol as described under ``Methods.'' Cells were then incubated with DMEM containing 1 mg/ml BSA and the indicated concentrations of HDL protein alone (, Control), with 4 µM brefeldin A (, BFA) or 25 µM monensin (, Mon) for 24 h at 37 °C. After incubation, efflux medium was collected, and cells were treated with cholesterol oxidase as described under ``Methods.'' Cell lipids were extracted and separated by TLC to isolate cholesterol, cholestenone (the cholesterol oxidase product), and cholesterol esters. Results were calculated as the percent of H in each fraction relative total H. A, efflux. B, oxidase-resistant cholesterol. C, oxidase-accessible cholesterol. D, cholesterol ester. Results are the means of duplicate incubations, representative of at least three experiments for each compound tested. Total H recovered was (mean ± S.D., n = 21) 55,137 ± 2384 cpm/dish, and there were no differences between groups.

Since these agents both affected cell [H]cholesterol esters, although in opposite directions, we examined whether the observed effects on efflux and cell distribution were due to these changes. A similar experiment as above was conducted in the presence of an ACAT inhibitor (Table 3). Monensin and brefeldin A induced qualitatively similar changes in cell distribution and HDL-mediated efflux of cholesterol under these conditions. Inhibition of ACAT increased cholesterol oxidase-accessible [H]cholesterol, with no appreciable change in oxidase-resistant cholesterol in control cells. Addition of monensin increased oxidase-resistant cholesterol similar to results without ACAT inhibition, whereas a small increase (from 10 to 14%) occurred in brefeldin A-treated cells that was not apparent without ACAT inhibition. These data suggest that inhibition of cholesterol efflux to HDL by these agents was independent of changes in the cholesterol ester content of cells.

Effects of monensin, nigericin, and brefeldin A on cholesterol efflux to HDL and cholesterol oxidase sensitivity were tested in porcine smooth muscle cells. Interestingly, and different from what was observed in fibroblasts, monensin and nigericin significantly increased efflux from cells to HDL-free medium (from 2.1% for controls to 9.4 and 10.5% for monensin and nigericin), although the reason for this has not been explored further. Efflux of cell [H]cholesterol to HDL was significantly inhibited (by approximately 40%) by all three compounds, and changes in cell cholesterol distribution were similar to those observed in fibroblasts for monensin and nigericin (data not shown). Brefeldin A increased oxidase-resistant and esterified cholesterol with a concomitant decrease in oxidase-accessible cholesterol in smooth muscle cells (data not shown), although the increase in oxidase-resistant sterol was less than observed for monensin and nigericin. Thus, the effects of these drugs on HDL-mediated cholesterol efflux and cell distribution were not identical in smooth muscle cells and fibroblasts and several similarities were noted, suggesting that these drugs exert their effects through common mechanisms in both experimental models.

The time course of cell [H]cholesterol efflux and cholesterol oxidase sensitivity of in the presence of absence of monensin and brefeldin A were examined ( Fig. 6and Fig. 7). In both studies, control dishes incubated without HDL showed no appreciable efflux of [H]cholesterol to medium (less than 2% of total cell radioactivity) and relatively no change among the various pools of cholesterol over the times examined, suggesting that these pools were in isotopic equilibrium. Efflux of cholesterol to medium containing HDL occurred in a time-dependent manner ( Fig. 6and Fig. 7). The majority of [H]cholesterol efflux was accounted for by depletion of cholesterol oxidase-accessible [H]cholesterol, especially at earlier times, and to depletion of cell [H]cholesterol esters, most obvious after 16 h. In these studies, similar to data in Fig. 5, HDL did not influence the levels of oxidase-resistant [H]cholesterol in control incubations. When monensin was present, [H]cholesterol efflux to HDL was similar to controls after 2 h, but decreased at all other times (Fig. 6). Inhibition of HDL-mediated efflux was maximal by 6 h (39% inhibition compared with controls) and similarly inhibited after 8 and 16 h (38 and 39%, respectively). Monensin decreased cholesterol oxidase-accessible sterol at all times, attaining a new basal level after 6 h (coincident with the maximal decrease in HDL-mediated efflux in monensin-treated cells) and paralleled by an increase in oxidase resistant [H]cholesterol. [H]Cholesterol esters decreased over time, relative to controls, also contributing to the increase in oxidase-resistant [H]cholesterol. Similar changes were observed when HDL and monensin were present together, except for a greater decrease in the cholesterol oxidase-accessible pool accounted for by [H]cholesterol appearing in medium. HDL was unable to decrease [H]cholesterol esters in monensin-treated cells. As observed previously, efflux of cell cholesterol to HDL in the presence of monensin was limited to removal of plasma membrane (i.e. cholesterol oxidase-accessible) cholesterol.

Figure 6: Effect of monensin on the time course of HDL-mediated efflux and cell distribution of [H]cholesterol in cholesterol-loaded human skin fibroblasts. Fibroblast cultures were labeled and enriched with cholesterol as described in the legend to Fig. 5. Cultures were incubated with DMEM containing 1 mg/ml BSA alone (), 50 µg/ml HDL (), 25 µM monensin (), or 25 µM monensin and 50 µg/ml HDL () for the indicated times. After incubation, medium was collected and cells were treated with cholesterol oxidase and results calculated as described in the legend to Fig. 5. A, efflux. B, oxidase-resistant cholesterol. C, oxidase-accessible cholesterol. D, cholesterol esters. Results are the means of duplicate incubations, representative of at least three experiments for each compound tested. Total H recovered was (mean ± S.D., n = 42) 31,485 ± 1,401 cpm/dish, and there were no differences between groups.

Figure 7: Effect of brefeldin A on the time course of HDL-mediated efflux and cell distribution of [H]cholesterol in cholesterol-loaded human skin fibroblasts. Fibroblast cultures were labeled and enriched with cholesterol as described in the legend to Fig. 5. Cultures were incubated with DMEM containing 1 mg/ml BSA alone (), 50 µg/ml HDL (), 4 µM brefeldin A (), or 4 µM brefeldin A and 50 µg/ml HDL () for the indicated times. After incubation, medium was collected, and cells were treated with cholesterol oxidase and results calculated as described in the legend to Fig. 6. A, efflux. B, oxidase-resistant cholesterol. C, oxidase-accessible cholesterol. D, cholesterol esters. Results are the means of duplicate incubations, representative of at least three experiments for each compound tested. Total H recovered was (mean ± S.D., n = 42) 56,453 ± 2,847 cpm/dish, and there were no differences between groups.

When brefeldin A was included during incubation with cells (Fig. 7), [H]cholesterol efflux to medium containing HDL was not appreciably different from controls until 6 h of incubation (12% inhibition); approximately 30% inhibition was seen at 6 and 8 h, and maximum inhibition of HDL-mediated efflux occurred at 16 h (46% decrease compared with control). The effects of brefeldin A on cholesterol oxidase sensitivity of cell [H]cholesterol was again different from changes observed with monensin. Brefeldin A caused a transient decrease in cholesterol oxidase accessible sterol, paralleled by a rise in the oxidase-resistant pool, which returned to near control levels by 8 h. This trend was not apparent in the presence of brefeldin A and HDL together, which did not demonstrate any measurable differences relative to control incubations containing HDL in either the cholesterol oxidase-accessible or -resistant pools. These data also demonstrate that brefeldin A did not alter efflux of plasma membrane (oxidase-accessible) cholesterol to HDL. In brefeldin A-treated cells [H]cholesterol esters were slightly increased over time relative to controls, apparent after 6 h and continuing to increase during the course of the experiment. When HDL was also present, this increase was attenuated, but when compared with control incubations with HDL, brefeldin A blocked HDL-mediated depletion of [H]cholesterol esters.

Cholesterol Mass Efflux

DISCUSSION

Monensin, nigericin, and brefeldin A blocked HDL-mediated efflux of intracellular cholesterol based on the following observations. First, these compounds prevented HDL from depleting the substrate pool of cholesterol available for esterification by ACAT. Since the enzyme resides within the endoplasmic reticulum(33) , the cholesterol substrate must be available to the same intracellular site. Second, monensin and brefeldin A prevented HDL to increase sterol biosynthesis compared with control cells. Third, treatment with monensin, nigericin, and brefeldin A inhibited HDL-mediated efflux of radiolabeled cholesterol and cholesterol mass from cholesterol-enriched cells.

We propose the following model for cholesterol flux through the cell (Fig. 8), making the following assumptions. First, cholesterol esters and oxidase-resistant cholesterol reside within intracellular compartments. Second, oxidase-sensitive cholesterol resides in the plasma membrane. Third, cholesterol esters participate in continuous hydrolysis and re-esterification (i.e. the cholesterol ester cycle). Fourth, cholesterol removed from cells is unesterified and passes through the plasma membrane before removal by an extracellular acceptor. In control cells overloaded with cholesterol, and incubated without an exogenous cholesterol or acceptors, a steady state is attained in which cholesterol oxidase-sensitive, -resistant and -esterified cholesterol remain at constant levels. These pools may be in equilibrium, and exchange among these pools may occur (indicated by reversible arrows, Fig. 8A). Results in cells incubated with an ACAT inhibitor indicate that hydrolyzed cholesterol esters transport to the plasma membrane without accumulating in the oxidase-resistant pool. When HDL is present, efflux is due to removal of plasma membrane (pathway a) and esterified cholesterol pools (pathway b). Decreases in cell cholesterol esters occur after hydrolysis, followed by transport and then uptake of free cholesterol from the plasma membrane. The oxidase-resistant cholesterol pool remains unchanged after incubation with HDL, suggesting that this pool of cholesterol is not available for efflux to HDL or, if depleted, is rapidly replenished by other cholesterol pools. Cholesterol ester hydrolysis does not cause the accumulation of free cholesterol in the oxidase-resistant or -sensitive pools, suggesting that transport of hydrolyzed cholesterol, once stimulated by appropriate extracellular acceptors, is not rate-limiting and rapidly removed from cells.

Figure 8: Pathways of cellular cholesterol transport and efflux. Potential pathways involved in cellular cholesterol transport and efflux to HDL in cholesterol-loaded cells and effects of monensin and brefeldin A. A, control conditions. B, monensin-treated cells. C, brefeldin A-treated cells. C, unesterified cholesterol; CE, cholesterol esters.

Monensin treatment of cells caused a redistribution of cell cholesterol (Fig. 8B). Oxidase-sensitive and -esterified cholesterol pools decrease, resulting in an increase in the cholesterol oxidase-resistant pool, and all pools attain an apparent new steady state over time. Re-distribution of cell cholesterol by monensin had little or no effect on the extent of plasma membrane cholesterol depletion by HDL (pathway a), in spite of a decrease in the size of this pool. In contrast, monensin blocks the ability of HDL to deplete intracellular cholesterol ester pools (pathway b), and similar to control conditions, the oxidase-resistant cholesterol pool is not available for efflux to HDL. Thus, monensin causes the redistribution of cholesterol into intercellular compartments and blocks transport to sites available for efflux to HDL.

Brefeldin A treatment of cells had different effects on cell cholesterol distribution than monensin. This compound caused a slight increase in the cell cholesterol ester pool due to a decrease in the oxidase-sensitive cholesterol pool (most notable in cholesterol-depleted cells). Efflux of plasma membrane cholesterol to HDL (pathway a) was comparable with control conditions. However, HDL could not deplete cholesterol ester pools (pathway b) in brefeldin A-treated cells, suggesting that transport of hydrolyzed cholesterol is blocked, and this pool of cholesterol was then efficiently re-esterified. Cholesterol ester hydrolysis rates were the same in brefeldin A-treated and controls cells, ()implying that changes in cholesterol ester hydrolysis did not cause the observed effects.

Based on this model, we suggest that HDL promotes cholesterol efflux from cells by two distinct pathways. Efflux of cholesterol from the plasma membrane (oxidase-sensitive) pool probably occurs by desorption and diffusion of cholesterol already present in this compartment, and some of the efflux from this pool may represent exchange between the cell and lipoprotein(2, 7) . Cholesterol efflux from cells by this mechanism does not depend on a functional and intact Golgi apparatus. A second pathway must also exist for the transport of intracellular cholesterol, derived from the hydrolysis of cholesterol esters, by a pathway that requires an intact and functional Golgi apparatus, revealed by sensitivity to monensin and brefeldin A. This pathway promotes cholesterol efflux in addition to efflux from the plasma membrane, and transport from intracellular sites to extracellular acceptors is rapid without causing the accumulation of cholesterol within any cellular pools.

Monensin affects the trans-cisternae of the Golgi apparatus in those regions primarily associated with the final stages of secretory vesicle maturation and in post-Golgi structures associated with endocytosis and membrane/product sorting (29) and preventing secretory vesicle production(33) . These effects of monensin have been used as one criterion for verifying passage of molecules through the Golgi apparatus ( (29) and references therein). These processes may account for the observed effects of monensin on cell cholesterol transport. Thus, monensin treatment of cells may prevent antegrade transport of cholesterol through the Golgi apparatus for delivery to the plasma membrane. Additionally, monensin may prevent cholesterol transport from the trans-Golgi region into the endoplasmic reticulum, effectively blocking cholesterol entry into the cholesteryl ester cycle. Retrograde transport of plasma membrane cholesterol to other intracellular sites may continue, but might accumulate at those sites if antegrade transport back to the plasma membrane depends on a functional trans-Golgi network.

Brefeldin A causes the disassembly of the Golgi apparatus, primarily the cis- and medial Golgi cisternae(30) , in contrast to monensin that primarily affects the trans-Golgi cisternae. The cis- and medial Golgi membranes redistribute with the endoplasmic reticulum, whereas components of the trans-Golgi network do not(30) . Retrograde transport of cis- and medial Golgi membrane back to the endoplasmic reticulum by the action of brefeldin A would maintain or increase substrate pools of cholesterol available for esterification. Disassembly of the Golgi apparatus would effectively block antegrade transport of cholesterol derived from hydrolysis of esters, resulting in the inability of cells to become depleted of intracellular cholesterol pools by efflux to an acceptor particle. However, cholesterol transport into cells and back to the plasma membrane at sites distal to the cis- and medial Golgi apparatus may not be affected by brefeldin A.

Previous studies have examined the effects of brefeldin A on cholesterol homeostasis in cultured cells. Stein et al.(34) showed that brefeldin A increased cholesterol esterification rates in cultured cells, attributed to increased substrate availability to sites of esterification due to the collapse of the Golgi apparatus into the endoplasmic reticulum without a direct effect on ACAT activity. Results from Hasumi et al.(35) using cultured macrophage cell lines showed that brefeldin A increased cell cholesteryl esters, at the expense of free cholesterol, but without affecting sterol biosynthetic rates. These authors reported that brefeldin A did not alter cholesterol oxidase sensitivity of cell cholesterol, however, without showing data. Neither of these studies examined the effects of brefeldin A on cholesterol transport to extracellular acceptors. More recently, Azhar et al.(37) demonstrated that okadaic acid prevented steroid hormone production from cholesterol in cultured cells and suggested this could result from the effect of this compound on disruption of the Golgi complex structure, implicating this organelle in providing cholesterol substrate to the mitochondria for steroid hormone production. Simoni and colleagues examined the effects of monensin (37) and brefeldin A (21) on the transport of newly synthesized cholesterol from intracellular sites of biosynthesis to the plasma membrane in Chinese hamster ovary cells. These studies did not demonstrate any effect on cholesterol transport by these compounds. Although these results were different to those of the present study, this may be accounted for by different pathways involved in transport of newly synthesized cholesterol in growing, non-cholesterol-loaded cells compared with transport of excess cholesterol in quiescent, cholesterol-loaded cells. These authors also showed that newly synthesized cholesterol was transported by a unique class of low density vesicles, apparently a post-endoplasmic reticulum intermediate (21, 37) , consistent with the idea that newly synthesized sterols are transported by novel vesicular transport pathways that bypass the Golgi apparatus. An additional caveat to those studies is that most of the newly synthesized sterol labeled during short pulse incubations are not cholesterol, but more polar sterol precursors(36) . Whether such molecules are transported in an identical manner as authentic cholesterol has not yet been addressed in that experimental system.

Cholesterol has been shown to accumulate in membranes enriched with sphingomyelin(39, 40) . As suggested by Shiao and Vance(41) , one possibility is that these lipids travel together in specialized vesicles to the plasma membrane. Newly synthesized sphingomyelin transport to the plasma membrane of hepatocytes was not affected by brefeldin A or monensin(41) , similar to effects of these compounds on transport of newly synthesized sterols(21, 37) . Brefeldin A also failed to prevent sphingomyelin transport in CaCO cells, although a redistribution of sphingomyelin between the apical and basolateral membranes did occur(42) . Based on those results and the present data, one would conclude that sphingomyelin and newly synthesized cholesterol are transported by Golgi-independent pathways (21, 37, 41, 42) distinct from transport of excess intracellular cholesterol that appears to be dependent on a functional Golgi complex.

In contrast, other studies have shown that brefeldin A and monensin block cellular transport of newly synthesized sphingomyelin(43, 44, 45) . Brefeldin A had no effects on the steady state levels of sphingomyelin mass, but reduced the proportion of plasma membrane sphingomyelin by 25% in baby hamster kidney cells(43) . In the same cell type, monensin stimulated degradation of plasma membrane sphingolipid and inhibited transport of newly synthesized sphingomyelin to the plasma membrane (44) . Monensin also inhibited transport of a fluorescent sphingolipid analog to the plasma membrane of Chinese hamster ovary cells(45) . Thus, depletion of plasma membrane sphingomyelin by monensin or brefeldin A may stimulate the internalization of membrane cholesterol. This concept is supported by studies of Slotte et al. (46, 47) demonstrating that degradation of plasma membrane sphingomyelin with exogenous sphingomyelinase promoted transport of cholesterol to intracellular compartments. These results suggest that plasma membrane sphingomyelin content directly influences cholesterol distribution. Internalization or degradation of plasma membrane sphingomyelin and intracellular accumulation, induced by brefeldin A or monensin, may result in the sequestration of cholesterol within those sphingomyelin-enriched membranes, and such effects may occur even if transport of these lipids occurs by different pathways. The relationship between sphingomyelin and cholesterol transport, especially under conditions of excess cholesterol accumulation and removal, deserve further investigation.

Pathways of intracellular cholesterol transport and their role in removal of excess cholesterol to extracellular acceptors remain incompletely understood. We have shown that agents that disrupt Golgi apparatus structure and function prevent efficient cholesterol removal by HDL and alter cellular cholesterol distribution. Such evidence strongly suggests a role for the Golgi apparatus in maintaining cell cholesterol distribution and in transport from intracellular sites to the plasma membrane for eventual removal. These studies did not directly address the mechanisms by which HDL promotes cholesterol efflux, such as by desorption and diffusion events (2, 3) or interactions with cell surface binding sites(13, 17, 18) , instead were designed to explore cellular pathways involved in cholesterol transport during efflux to an appropriate cholesterol acceptor, in this case HDL. The Golgi apparatus is composed of multiple subcompartments with continuous membrane exchange occurring between them(48) . The Golgi apparatus contains substantial amounts of cholesterol, and there is evidence for a cholesterol gradient in the cis to trans direction(49) . One may speculate that cholesterol is transported from the cis- to trans-Golgi (or antegrade transport) and possibly accumulates within trans-Golgi vesicles until the plasma membrane can accommodate more cholesterol, such as after depletion of plasma membrane cholesterol by appropriate acceptors or stimulation by signaling molecules(17, 18, 19, 20) . Alternatively, there may be continuous recycling of cholesterol between the plasma membrane and the Golgi complex, such that a ``transport equilibrium'' exists between these two membrane systems and possibly between the Golgi apparatus and other intracellular compartments such as the endoplasmic reticulum or mitochondria. Removal of cholesterol from the plasma membrane by acceptors would shift the equilibrium to deliver more cholesterol from Golgi membranes to the plasma membrane. Influx of cholesterol, which initially accumulates in the plasma membrane(10, 11, 12) , would shift the equilibrium to deliver more cholesterol to the Golgi apparatus and eventually transport cholesterol to the endoplasmic reticulum for esterification and storage. In either scheme the Golgi apparatus would be involved in various aspects of cellular cholesterol transport and targeting, not only to the plasma membrane but also to other sites of cholesterol utilization or storage. Further studies are needed to examine the possible steps in Golgi-mediated intracellular cholesterol transport, and elucidation of these pathways will increase our understanding of the mechanisms involved in excess cholesterol accumulation and removal from cells.

FOOTNOTES

*

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore by hereby marked ``advertisement'' in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

§

To whom correspondence should be addressed: Massachusetts General Hospital, Cardiac Unit, Jackson 1422, 32 Fruit Street, Boston, MA 02114. Tel.: 617-726-3757; Fax: 617-726-4811.

()

The abbreviations used are: HDL, high density lipoproteins; ACAT, acyl coenzyme A:cholesterol acyltransferase; BSA, fatty acid free bovine serum albumin; DMEM, Dulbecco's modified Eagle's medium; FBS, fetal bovine serum; LDL, low density lipoproteins; LPDS, lipoprotein-deficient fetal bovine serum; PBS, phosphate-buffered saline.

()

A. J. Mendez, unpublished observations.

ACKNOWLEDGEMENTS

REFERENCES

1. Gordon, D. J., Probstfield, J. L., Garrison, R. J., Neaton, J. D., Castelli, W. P., Knoke, J. D., Jacobs, D. R., Bangdiwala, S., and Tyroler, H. F. (1989) Circulation 79, 5-15
2. Phillips, M. C., Johnson, W. J., and Rothblat, G. H. (1987) Biochim. Biophys. Acta 906, 223-276 [Medline] [Order article via Infotrieve]
3. Johnson, W. J., Mahlberg, F. H., Rothblat, G. H., and Phillips, M. C. (1991) Biochim. Biophys. Acta 1085, 273-298 [Medline] [Order article via Infotrieve]
4. Fielding, C. J., and Fielding, P. E. (1982) Med. Clin. N. Am. 66, 363-373 [Medline] [Order article via Infotrieve]
5. DeLamatre, J., Wolfbauer, G., Phillips, M. C., and Rothblat, G. H. (1986) Biochim. Biophys. Acta 875, 419-428 [Medline] [Order article via Infotrieve]
6. Barkia, A., Puchois, P. Nordine, G., Torpier, G., Barbaras, R., Ailhaud, G., and Fruchart, J.-C. (1991) Atherosclerosis 87, 135-146 [CrossRef][Medline] [Order article via Infotrieve]
7. Rothblat, G. H., Mahlberg, F. H., Johnson, W. J., and Phillips, M. C. (1992) J. Lipid Res. 33, 1091-1087 [Abstract]
8. Mahlberg, F. H., and Rothblat, G. H. (1992) J. Biol. Chem. 267, 4541-4550 [Abstract/Free Full Text]
9. Brown, M. S., and Goldstein, J. L. (1986) Science 232, 34-47 [Free Full Text]
10. Tabas, I., Rosoff, W. J., and Boykow, G. C. (1988) J. Biol. Chem. 263, 1266-1272 [Abstract/Free Full Text]
11. Lange, Y., Strebel, F., and Steck, T. L. (1993) J. Biol. Chem. 268, 13838-13843 [Abstract/Free Full Text]
12. Johnson, W. J., Chacko, G. K., Phillips, M. C., and Rothblat, G. H. (1990) J. Biol. Chem. 265, 5546-5553 [Abstract/Free Full Text]
13. Slotte, J. P., Oram, J. F., and Bierman, E. L. (1987) J. Biol. Chem. 262, 12904-12907 [Abstract/Free Full Text]
14. Oram, J. F., Mendez, A. J., Slotte, J. P., and Johnson, T. J. (1991) Arterioscler. Thromb. 11, 403-414 [Abstract/Free Full Text]
15. Mendez, A. J., Anantharamaiah, G. M., Segrest, J., and Oram, J. F. (1994) J. Clin. Invest. 94, 1698-1705
16. Lee, M., Lindstedt, L. K., and Kovanen, P. T. (1992) Arterioscler. Thromb. 12, 1329-1335 [Abstract/Free Full Text]
17. Therit, N., Delbart, C., Aguie, G., Fruchart, J.-C., Vassaux, G., and Ailhaud, G. (1990) Biochim. Biophys. Res. Commun. 173, 1361-1368 [CrossRef][Medline] [Order article via Infotrieve]
18. Mendez, A. J., Oram, J. F., and Bierman, E. L. (1991) J. Biol. Chem. 266, 10104-10111 [Abstract/Free Full Text]
19. Hokland, B. M., Slotte, J. P., Bierman, E. L., and Oram, J. F. (1994) J. Biol. Chem. 268, 25343-25349 [Abstract/Free Full Text]
20. Bernard, D. W., Rodriguez, A., Rothblat, G. H., and Glick, J. M. (1991) J. Biol. Chem. 266, 710-716 [Abstract/Free Full Text]
21. Urbani, L., and Simoni, R. D. (1990) J. Biol. Chem. 265, 1919-1923 [Abstract/Free Full Text]
22. Ross, R. (1971) J. Cell Biol. 50, 172-186 [Abstract/Free Full Text]
23. Wiesgraber, K. H., and Mahley, R. W. (1980) J. Lipid Res. 21, 623-634
24. Lowry, O. H., Rosebrough, N., Farr, A., and Randall, R. J. (1951) J. Biol. Chem. 193, 265-275 [Free Full Text]
25. Oram, J. F. (1986) Methods Enzymol. 129, 645-659 [Medline] [Order article via Infotrieve]
26. Folch, J., Lees, M., and Sloane Stanley, G. (1957) J. Biol. Chem. 226, 497-509 [Free Full Text]
27. Lange, Y., and Ramos, B. V. (1983) J. Biol. Chem. 258, 15130-15134 [Abstract/Free Full Text]
28. Cabaud, P. G., and Wrobblewski, F. (1958) Am. J. Clin. Pathol. 30, 324-329
29. Mollenhauer, H. H., Morré, D. J., and Rowe, L. D. (1990) Biochim. Biophys. Acta 1031, 225-246 [Medline] [Order article via Infotrieve]
30. Klausner, R. D., Donaldson, J. G., and Lippincott-Schwartz, J. (1992) J. Cell Biol. 116, 1071-1080 [Free Full Text]
31. Basu, S. K., Goldstein, J. L., Anderson, R. G. W., and Brown, M. S. (1981) Cell 24, 493-502 [CrossRef][Medline] [Order article via Infotrieve]
32. Oram, J. F. (1983) Arteriosclerosis 3, 420-432 [Abstract/Free Full Text]
33. Morré, D. J., Boss, W. F., Grimes, H., and Mollenhauer, H. H. (1983) Eur. J. Cell Biol. 30, 25-32 [Medline] [Order article via Infotrieve]
34. Stein, O., Dabach, Y., Hollander, G., Ben-Nam, M., and Stein, Y. (1992) Biochim. Biophys. Acta 1125, 28-34 [Medline] [Order article via Infotrieve]
35. Hasumi, K., Naganuma, S., Koshizawa, J., Mogi, H., and Endo, A. (1993) Biochim. Biophys. Acta 1167, 155-158 [Medline] [Order article via Infotrieve]
36. Azhar, S., Frazier, J. A., Tsai, L., and Reaven, E. (1994) J. Lipid Res. 35, 1161-1176 [Abstract]
37. Kaplan, M. R., and Simoni, R. D. (1985) J. Cell Biol. 101, 446-453 [Abstract/Free Full Text]
38. Echevarria, F., Norton, R. A., Nes, W. D., and Lange, Y. (1990) J. Biol. Chem. 265, 8484-8489 [Abstract/Free Full Text]
39. Wattenberg, B. W., and Silbert, D. F. (1983) J. Biol. Chem. 258, 2284-2289 [Free Full Text]
40. Yeagle, P. L., and Young, J. E. (1986) J. Biol. Chem. 261, 8175-8181 [Abstract/Free Full Text]
41. Shiao, Y.-J., and Vance, J. E. (1993) J. Biol. Chem. 268, 26085-26092 [Abstract/Free Full Text]
42. Van Meer, G., and Van't Hof (1993) J. Cell. Sci. 104, 833-842 [Abstract]
43. Kallen, K.-J., Quinn, P., and Allan, D. (1993) Biochem. J. 289, 307-312
44. Kallen, K.-J., Quinn, P., and Allan, D. (1993) Biochim. Biophys. Acta 1166, 305-308 [Medline] [Order article via Infotrieve]
45. Lipsky, N. G., and Pagano, R. E. (1985) J. Cell Biol. 100, 27-34 [Abstract/Free Full Text]
46. Slotte, J. P., and Bierman, E. L. (1988) Biochem. J. 250, 653-658 [Medline] [Order article via Infotrieve]
47. Slotte, J. P., Tenhyunen, J., and Pörn, I. (1990) Biochim. Biophys. Acta 1025, 152-156 [Medline] [Order article via Infotrieve]
48. Mellman, I., and Simons, K. (1992) Cell 68, 829-840 [CrossRef][Medline] [Order article via Infotrieve]
49. Orci, L., Montesano, R., Meda, P., Malaisse-Lagae, F., Brown, D., Perrelet, A., and Vassali, P. (1981) Proc. Natl. Acad. Sci. U. S. A. 78, 293-297 [Abstract/Free Full Text]

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				T. D. Southall, S. Terhzaz, P. Cabrero, V. R. Chintapalli, J. M. Evans, J. A. T. Dow, and S.-A. Davies Novel subcellular locations and functions for secretory pathway Ca2+/Mn2+-ATPases Physiol Genomics, September 14, 2006; 26(1): 35 - 45. [Abstract] [Full Text] [PDF]

				H. Zheng, R. S. Kiss, V. Franklin, M.-D. Wang, B. Haidar, and Y. L. Marcel ApoA-I Lipidation in Primary Mouse Hepatocytes: SEPARATE CONTROLS FOR PHOSPHOLIPID AND CHOLESTEROL TRANSFERS J. Biol. Chem., June 3, 2005; 280(22): 21612 - 21621. [Abstract] [Full Text] [PDF]

				P. Matarrese, A. Tinari, E. Mormone, G. A. Bianco, M. A. Toscano, B. Ascione, G. A. Rabinovich, and W. Malorni Galectin-1 Sensitizes Resting Human T Lymphocytes to Fas (CD95)-mediated Cell Death via Mitochondrial Hyperpolarization, Budding, and Fission J. Biol. Chem., February 25, 2005; 280(8): 6969 - 6985. [Abstract] [Full Text] [PDF]