Mechanisms of high density lipoprotein-mediated efflux of cholesterol from cell plasma membranes

Size: px

Start display at page:

Download "Mechanisms of high density lipoprotein-mediated efflux of cholesterol from cell plasma membranes"

Elisabeth Hardy
6 years ago
Views:

Atherosclerosis 137 Suppl. (1998) S13 S17 Mechanisms of high density lipoprotein-mediated efflux of cholesterol from cell plasma membranes Michael C. Phillips *, Kristin L. Gillotte, M.

1 Atherosclerosis 137 Suppl. (1998) S13 S17 Mechanisms of high density lipoprotein-mediated efflux of cholesterol from cell plasma membranes Michael C. Phillips *, Kristin L. Gillotte, M. Page Haynes, William J. Johnson, Sissel Lund-Katz, George H. Rothblat Biochemistry Department, MCP-Hahnemann School of Medicine, Allegheny Uni ersity of the Health Sciences, 2900 Queen Lane, Philadelphia, PA 19129, USA Abstract The participation of HDL in the reverse cholesterol transport (RCT) from peripheral cells to the liver is critical for the antiatherogenic properties of this lipoprotein. Experimental results showing that efflux of cholesterol from cells growing in culture is mediated by HDL and lipoprotein particles containing apo A-I, in particular, support this conclusion. A bidirectional flux of unesterified cholesterol molecules between the plasma membrane of cells and HDL particles in the extracellular medium occurs. Net efflux of cholesterol mass from the cells involves passive diffusion of cholesterol molecules through the aqueous phase and down their concentration gradient between the membrane and HDL; the concentration gradient is maintained by LCAT-mediated esterification of cholesterol molecules in the HDL particles. Fully lipidated apo A-I is important in promoting this aqueous diffusion mechanism because it: (1) acts as a cofactor for LCAT; and (2) solubilizes phospholipid into small HDL-sized particles that are efficient at absorbing cholesterol molecules diffusing away from the cell surface. Apo A-I also exists in an incompletely lipidated state in plasma. Apo A-I molecules in this state are able to solubilize phospholipid and cholesterol from the plasma membrane of cells. This membrane-microsolubilization process is enhanced by enrichment of the plasma membrane with cholesterol and is the mechanism by which pre- -HDL particles in the extracellular medium remove cholesterol and phospholipid from cells. The relative contributions in vivo of the aqueous diffusion and membrane-microsolubilization mechanisms of apo A-I-mediated cell cholesterol efflux are not predicted readily from cell culture experiments. Confounding issues are the variations with cell type and the dependence on the degree of cholesterol loading of the cell plasma membrane Elsevier Science Ireland Ltd. All rights reserved. Keywords: Cholesterol; High density lipoprotein; Plasma membrane 1. Introduction A major function of high density lipoprotein (HDL) is the transport of cholesterol from peripheral cells to the liver where the cholesterol can be converted to bile acids and excreted from the body [1]. This reverse cholesterol transport (RCT) pathway is crucial for cholesterol homeostasis in peripheral cells and the promotion of this transport by HDL underlies the antiatherogenicity of this lipoprotein [2,3]. The first step in RCT involves efflux of cholesterol from the plasma * Corresponding author. Tel.: ; fax: membrane of peripheral cells to HDL in the extracellular medium. There have been many studies of this step over the past 20 years examining both the mechanism and the influence of the HDL structure [1,4 7]. There is general agreement that there are two mechanisms by which HDL and its apolipoprotein can induce cell cholesterol efflux. Apo A-I is the principal protein of HDL and, as such, is a major antiatherogenic protein. This property of apo A-I has been demonstrated directly by the reduced incidence of atherosclerotic plaque in mice overexpressing human apo A-I [8]. The role of apo A-I in mediating efflux of cholesterol from cell plasma membranes depends upon its state of lipidation. Fully /98/$ Elsevier Science Ireland Ltd. All rights reserved. PII S (97)

2 S14 M.C. Phillips et al. / Atherosclerosis 137 Suppl. (1998) S13 S17 Fig. 1. Molecular events involved in efflux of cholesterol (C) from the plasma membrane of peripheral cells. The scheme depicts the aqueous diffusion and membrane-microsolubilization mechanisms by which lipidated and lipid-free (poor) forms of apo A-I remove C. The esterification of HDL cholesterol by LCAT reduces the possibility of cholesterol molecules diffusing back to the cell thereby promoting the RCT pathway. See the text for further details. lipidated apo A-I molecules as present in mature HDL 2 and HDL 3 particles remove cholesterol from cells by the aqueous diffusion mechanism. On the other hand, lipid-free or lipid-poor apo A-I as exists in pre- -HDL induces cell cholesterol efflux by a membrane-microsolubilization process. Fig. 1 gives a schematic overview of these two processes. It is apparent that apo A-I is involved in the following key interactions: (1) with the lipids of HDL particles as a structural component; (2) with the surface of the plasma membrane in the membrane-microsolubilization process; (3) with lecithin-cholesterol acyltransferase (LCAT) as a cofactor promoting esterification of cholesterol removed from cells by HDL [9]; and (4) at the end of the RCT pathway (not shown in Fig. 1), with scavenger receptor (SR) B-1 in the hepatocyte plasma membrane as a ligand mediating the selective uptake of cholesteryl ester from HDL particles [10]. In this article, we summarize current understanding of the aqueous diffusion and membrane-microsolubilization mechanisms. There are important outstanding issues. For instance, the relative contributions of the two pathways in vivo have not been established; it is likely that the local activity of LCAT is important in this regard with high activity favouring cholesterol efflux by aqueous diffusion rather than membrane-microsolubilization. In addition, the reasons for variations in behavior between cell types are not clear. It is likely that the domain structure of the plasma membrane [11] is important with domains being accessed differently by various HDL subspecies. These important issues remain to be resolved by future experiments. 2. Contributions of HDL components to cholesterol efflux The apolipoprotein and phospholipid (PL) components of HDL can mediate cell cholesterol efflux and Fig. 2 compares the functionalities of these constituents Fig. 2. Timecourses of [ 3 H]FC efflux from human skin fibroblasts to apo A-I, SUV and reconstituted HDL discoidal complexes. Normal human skin fibroblasts were grown to confluence in 36-mm tissue culture wells and trace labeled with [ 3 H]FC in 1% fetal bovine serum. The radiolabeled monolayers were incubated at 37 C with 2 ml of human apo A-I ( ), SUV ( ), or discoidal POPC:apo A-I (2.0:1 w/w) complexes ( ) for periods ranging from 0 to 6 h. Apo A-I and reconstituted HDL disks were applied at a concentration of 50 g protein/ml and the SUV acceptor was utilized at 100 g PL/ml, a concentration comparable to the amount of PL present in the HDL treatment. At specific timepoints an aliquot was removed, filtered and analyzed for [ 3 H]cholesterol content by liquid scintillation counting. Symbols represent the mean percent of FC release measured from triplicate wells 1 S.D. All values have been corrected for flux measured to minimal essential media in the absence of a protein or PL acceptor.

5 M.C. Phillips et al. / Atherosclerosis 137 Suppl. (1998) S13 S17 S15 in this regard. Lipid-free apo A-I induces relatively little efflux of cholesterol mass indicating that its capacity to retain cell cholesterol is low. On the other hand, the capacity of PL small unilamellar vesicles (SUV) is higher but cholesterol efflux is relatively slow; when 100 g/ml PL is present, about 2% of cell cholesterol is released in 6 h. The data in Fig. 2 demonstrate clearly that, when given amounts of apo A-I and PL are present together as a discoidal complex, efficient cholesterol efflux occurs. Thus the same concentration of apo A-I present as a preformed PL/apo A-I discoidal complex gives approximately ten times more efflux than the lipid-free protein. The same concentration of PL present as a disk gives approximately four times more efflux than when it is present as a small unilamellar vesicle. These results show that apo A-I promotes efflux by solubilizing the PL into disks containing approximately 200 PL molecules compared to SUV which contain some 2000 PL molecules per particle. The amphipathic -helical segments in the apo A-I molecule [12] interact with the PL molecules to stabilize the small disks which are efficient at mediating cell cholesterol efflux [13]. The mechanisms by which fully lipidated and lipid-free (poor) apo A-I induce cholesterol efflux from cells are considered in turn below. 3. Aqueous diffusion mechanism Mature HDL 2 and HDL 3 particles participate in a bidirectional flux of free cholesterol (FC) molecules between the lipoprotein and cells with the direction of net transfer of cholesterol mass being determined by the gradient in cholesterol concentration [1,13]. This exchange of FC occurs by a so-called aqueous diffusion mechanism in which cholesterol molecules desorb from the donor lipid-water interface and diffuse through the intervening aqueous layer until they collide with and are absorbed by an acceptor particle. The rate of cholesterol exchange or transfer is first order with respect to the concentration of FC in the donor particle and is strongly temperature-dependent. The rate constant for cholesterol transfer is about an order of magnitude greater than that for PL transfer [13]. Complex kinetics can be observed for the transfer of lipid molecules by diffusion, with the reaction order depending upon the nature and concentrations of the donor and acceptor particles. The rate of FC efflux from a monolayer of fibroblasts shows a hyperbolic dependence on the concentration of HDL acceptor particles in the extracellular medium (Fig. 3). The hyperbolic dependence of FC transfer rate on acceptor concentration is explained in terms of a kinetic scheme in which a first-order reaction to form a FC intermediate state is followed by a second-order interaction with the acceptor particle to transfer the FC molecule from the intermediate state to the acceptor particle. Fig. 3. Concentration-dependence of the rate constant for cholesterol efflux (k e ) from mouse L-cell fibroblasts to discoidal reconstituted HDL. Mouse L-cell fibroblasts grown to confluence in 22-mm tissue culture wells and trace labeled with [ 3 H]FC were incubated with 1 ml test medium, containing 0.5% bovine serum albumin and the indicated PL concentration of POPC:apo A-I (100:1, mol:mol) discoidal HDL. A 6-h timecourse of efflux was measured in triplicate at each concentration and fitted by nonlinear regression for derivation of the average FC efflux rate constant (k e ). Each point represents the mean k e and the error bars represent 1 S.D. The scheme for donor cells (D) and acceptor HDL particles (A) can be written in terms of the two steps [13,14] DC k 1 D+C k i (1) 1 A+C i k 2 AC (2) k 2 where DC and AC represent FC in the donor and acceptor particles, respectively. C i is FC in the intermediate state (monomer in the aqueous phase) and the k values are the rate constants for the steps indicated. Under initial velocity conditions where k 2 [AC] 0, the rate of formation of C i from mass action kinetics is given by Eq. (3). d[c i ] dt =k 1[DC] [C i ](k 1 [D]+k 2 [A]) (3) At the steady state, d[c i ]/dt=0 so that k [C 1 [DC] i ]= (4) (k 1 [D]+k 2 [A]) from Eq. (2), the initial velocity ( ) of the transfer reaction = d[ac] =k 2 [A][C i ]= (k 1k 2 [DC][A]) (5) dt (k 1 [D]+k 2 [A]) where the terms in square brackets are concentration at time t 0. Eq. (5) describes the hyperbolic dependence of on the acceptor particle concentration. When k 1 [D] k 2 [A], there is first-order dependence of on acceptor concentration because =k app [A] at constant donor concentration; the apparent rate constant k app = k 1 k 2 [DC]/k 1 [D] contains collisional rate constants k 1

7 S16 M.C. Phillips et al. / Atherosclerosis 137 Suppl. (1998) S13 S17 and k 2. When k 2 [A] k 1 [D], or simply [A] [D] when both the donor and acceptor species are similar so that similar on-rates for FC are expected (i.e. k 2 k 1 ), =k 1 [DC] which indicates that is independent of the acceptor concentration [A] and zero-order kinetics occur at constant donor concentration. The rate constant k 1 describes the diffusion of FC from the donor particle into the intermediate state C i (i.e. desorption into the aqueous phase); any factors that modulate the packing of FC molecules in the surface of the donor particle (cell plasma membrane) can influence the rate-limiting step k 1 [13]. The concentration of acceptor particles required to achieve the maximum rate of cholesterol efflux from cells depends upon the size of the acceptor particles [13,15]. For example, the maximum rate of efflux occurs with discoidal HDL particles present at 500 g PL/ml (Fig. 3), whereas, the concentration of PL present as SUV required for the maximum rate is much higher [15]. When FC efflux data for subsaturating concentrations of discoidal HDL and unilamellar PL vesicles are normalized on the basis of the number of acceptor particles present, large particles give more efficient efflux than small particles at the same particle number concentration [15]. Under this condition, there are more frequent collisions between desorbed cholesterol molecules and the acceptor particles. 4. Membrane-microsolubilization by apo A-I Approximately 5% of the apo A-I in human plasma exists in a lipid-free (poor) pool [5,16] which exhibits pre- mobility on agarose gel electrophoresis. This pre- pool of apo A-I is preferred over fully-lipidated apo A-I in spherical HDL particles as the acceptor of cell cholesterol when fibroblasts are incubated with plasma for 1 min [5]. As discussed above, the lipid-free (poor) apo A-I has a limited capacity to accept cholesterol; LCAT-mediated conversion to larger particles is required to enhance the capacity [5]. Free apo A-I (and other apolipoproteins) can remove FC and PL from cholesterol-loaded cells [6]. Fig. 4 depicts data for fibroblasts confirming that apo A-I can remove FC and PL together in a membranemicrosolubilization process. Furthermore, it is apparent that FC-loading of the fibroblasts enhances efflux of both FC and PL to apo A-I. The rates of apo A-I-mediated lipid efflux vary with cell type and are particularly rapid with cholesterol-loaded macrophages [6,17]. The increased cholesterol content of the cell apparently alters plasma membrane structure so that apo A-I can interact more readily and remove FC and PL molecules. The mechanism of this process is not known; the involvement of either particular membrane domains or membrane proteins and of particular regions of the apo A-I molecule requires elucidation. Cyclodextrins are a convenient tool for revealing pools of plasma membrane cholesterol that exhibit different efflux kinetics because the cyclodextrins promote extremely rapid efflux of cell cholesterol [18]. The results of such an experiment are shown in Fig. 5 and indicate that FC enrichment in fibroblasts does alter plasma membrane domain structure [11]. Doubling the FC content of the cells decreases the size of the pool exhibiting fast efflux (Fig. 5). This effect indicates that FC enrichment alters the distribution of cholesterol in the plasma membrane. Further work is required to define this structural change and learn what role it plays in the enhanced membrane-microsolubilization that occurs when apo A-I is incubated with cholesterolloaded cells. 5. Summary and conclusions It is clear that there are two pathways by which apo A-I can mediate efflux of cellular cholesterol. (1) HDL particles that contain fully lipidated apo A-I molecules participate in the aqueous diffusion process; this involves the desorption of cholesterol molecules from the cell plasma membrane and their diffusion through the aqueous phase where they can become absorbed by the HDL acceptor particles. (2) Incompletely lipidated apo A-I molecules in the lipid-free (poor) pre- -HDL pool Fig. 4. Cholesterol and PL efflux from control or FC-enriched human skin fibroblast monolayers to human apo A-I. Normal human skin fibroblasts were grown to confluence in 22-mm tissue culture wells and were trace-labeled for 2 days with [ 3 H]choline chloride and [ 14 C]FC in 10% human lipoprotein-deficient serum. The FC-enriched cells were prepared by an additional 24-h incubation with FC/PL liposomes and human LDL in the presence of an ACAT inhibitor. The prepared monolayers were incubated at 37 C with apo A-I in triplicate at 50 g/ml in minimal essential medium; at specific timepoints an aliquot of medium was removed, filtered, extracted and analyzed for [ 3 H]PL and [ 14 C]cholesterol content by liquid scintillation counting. The bars represent the mean percent 1 S.D. of FC and PL released from control (closed bars) or FC-enriched (hatched bars) monolayers during a 4 h incubation and all data have been corrected for sterol release to minimal essential medium in the absence of apo A-I. The FC contents of control and enriched cells were about g/mg cell protein, respectively.

8 M.C. Phillips et al. / Atherosclerosis 137 Suppl. (1998) S13 S17 S17 References Fig. 5. Timecourses of cholesterol efflux from control and FC-enriched human skin fibroblast monolayers to a 50 mm cyclodextrin solution. Normal human skin fibroblasts were grown to confluence in 22-mm tissue culture wells and were trace-labeled for 2 days with [ 3 H]FC in 10% human lipoprotein-deficient serum. FC-enriched cells were prepared by an additional 24-h incubation with FC/PL liposomes and human LDL in the presence of an ACAT inhibitor. The prepared monolayers were incubated at 37 C with 50 mm hydroxypropyl- cyclodextrin solution (50% saturated with cholesterol) in minimal essential medium; at specific timepoints an aliquot was removed, filtered and analyzed for [ 3 H]cholesterol content by liquid scintillation counting. Each point represents the mean fraction of [ 3 H]cholesterol remaining in the cells at the specified time (, FC-enriched fibroblasts;, control fibroblasts). The FC contents of control and enriched cells differed by a factor of about two. remove cholesterol and PL from the plasma membrane from cells by a membrane-microsolubilization process. The mechanism and in particular the nature of the apo A-I/plasma membrane interaction, remains to be resolved. It is likely that the relative contributions of these two pathways to cellular cholesterol efflux in vivo vary depending upon the local environment. Complicating issues are the cell type involved and the activity of LCAT in the extracellular space. High LCAT activity tends to reduce the size of the lipid-free pool of apo A-I in plasma [2,3,5], thereby, reducing the relative contribution of membrane microsolubilization (Fig. 1). Acknowledgements The research from this laboratory reported in this paper was supported by NIH grants HL07443 and HL [1] Johnson WJ, Mahlberg FH, Rothblat GH, Phillips MC. Cholesterol transport between cells and high density lipoproteins. Biochim Biophys Acta 1991;1085: [2] Barter PJ, Rye K-A. Molecular mechanisms of reverse cholesterol transport. Curr Opin Lipidol 1996;7:82 7. [3] Barter PJ, Rye K-A. High density lipoproteins and coronary heart disease. Atherosclerosis 1996;121:1 12. [4] Pieters MN, Schouten D, Van Berkel TJC. In vitro and in vivo evidence for the role of HDL in reverse cholesterol transport. Biochim Biophys Acta 1994;1225: [5] Fielding CJ, Fielding PE. Molecular physiology of reverse cholesterol transport. J Lipid Res 1995;36: [6] Oram JF, Yokoyama S. Apolipoprotein-mediated removal of cellular cholesterol and phospholipids. J Lipid Res 1996;37: [7] von Eckardstein A. Cholesterol efflux from macrophages and other cells. Curr Opin Lipidol 1996;7: [8] Schultz JR, Rubin EM. The properties of HDL in genetically engineered mice. Curr Opin Lipidol 1994;5: [9] Glomset JA. The plasma lecithin:cholesterol acyltransferase reaction. J Lipid Res 1968;9: [10] Acton S, Rigotti A, Landschulz KT, Xu S, Hobbs HH, Krieger M. Identification of scavenger receptor SR B-1 as a high density lipoprotein receptor. Science 1996;271: [11] Rothblat GH, Mahlberg FH, Johnson WJ, Phillips MC. Apolipoprotein, membrane cholesterol domains and the regulation of cholesterol efflux. J Lipid Res 1992;33: [12] Brouillette CG, Anantharamaiah GM. Structural models of human apo A-I. Biochim Biophys Acta 1995;1256: [13] Phillips MC, Johnson WJ, Rothblat GH. Mechanisms and consequences of cellular cholesterol exchange and transfer. Biochim Biophys Acta 1987;906: [14] Nichols JW, Pagano RE. Kinetics of soluble lipid monomer diffusion between vesicles. Biochemistry 1981;20: [15] Davidson WS, Rodrigueza WV, Lund-Katz S, Johnson WJ, Rothblat GH, Phillips MC. Effects of acceptor particle size on the efflux of cellular free cholesterol. J Biol Chem 1995;270: [16] Asztalos BF, Roheim PS. Presence and formation of free apo A-I-like particles in human plasma. Arterioscler Thromb Vasc Biol 1995;15: [17] Yancey PG, Bielicki JK, Johnson WJ, Lund-Katz S, Palgunachari MN, Anantharamaiah GM, Segrest JP, Phillips MC, Rothblat GH. The efflux of cellular cholesterol and phospholipid to lipid-free apolipoproteins and class A amphipathic peptides. Biochemistry 1995;34: [18] Yancey PG, Rodrigueza WV, Kilsdonk EPC, Stoudt GW, Johnson WJ, Phillips MC, Rothblat GH. Cellular cholesterol efflux mediated by cyclodextrins: demonstration of kinetic pools and mechanism of efflux. J Biol Chem 1996;271:

9 .

Removal of cellular cholesterol by pre- -HDL involves plasma membrane microsolubilization

Removal of cellular cholesterol by pre- -HDL involves plasma membrane microsolubilization Kristin L. Gillotte, 1 W. Sean Davidson, 2 Sissel Lund-Katz, George H. Rothblat, and Michael C. Phillips 3 Department