Abstract
Vascular smooth muscle cells (SMCs) grown in primary culture are converted from a contractile to a synthetic phenotype. This includes a marked morphological reorganization, with loss of myofilaments and formation of a large ER-Golgi complex. In addition, the number of cell surface caveolae is distinctly reduced and the handling of lipoprotein-derived cholesterol changed. Here we used filipin as a marker to study the distribution of cholesterol in SMCs by electron microscopy. In contractile cells, filipin-sterol complexes were preferentially found in caveolae and adjacent ER cisternae (present in both leaflets of the membranes). After exposure to LDL or cholesterol, labeling with filipin was increased both in membrane organelles and in the cytoplasm. In contrast, treatment with mevinolin (a cholesterol synthesis inhibitor) or β-cyclodextrin (a molecule that extracts cholesterol from cells) decreased the reaction with filipin but did not affect the close relation between the ER and the cell surface. In synthetic cells, filipin-sterol complexes were diffusely spread in the plasma membrane and the strongest cytoplasmic reaction was noted in endosomes/lysosomes, both under normal conditions and after incubation with LDL or cholesterol. On the basis of the present findings, we propose a mechanism for direct exchange of cholesterol between the plasma membrane and the ER and more active in contractile than in synthetic SMCs.
CAVEOLAE are small, flask-shaped invaginations of the plasma membrane with important functions in signal transduction (Anderson 1998; Smart et al. 1999) and lipid transport (Fielding and Fielding 1997; Ikonen 1997). They exist in most cell types and are particularly abundant in endothelial cells and smooth muscle cells (SMCs) of blood vessels. Biochemically, caveolae represent a plasma membrane subdomain enriched in cholesterol, glycosphingolipids, and the structural protein caveolin (three mammalian isoforms identified to date). Caveolin is organized as oligomers in the membrane and forms a scaffold on which a large variety of signaling molecules have been described to gather in preassembled signaling complexes (Smart et al. 1999). Furthermore, caveolin binds cholesterol directly (Murata et al. 1995) and is suggested to participate in the shuttling of free cholesterol between the endoplasmic reticulum (ER), the Golgi complex, and the cell surface (Conrad et al. 1995; Smart et al. 1996; Fielding and Fielding 1996; Uittenbogaard et al. 1998). Efflux of cellular free cholesterol likewise appears to be mediated via caveolae (Fielding and Fielding 1995).
We have previously reported that there is a phenotype-dependent variation in the number of caveolae in vascular SMCs both in vivo and in vitro. The differentiated contractile SMCs that are found in the normal arterial media and during the first 1–2 days of primary culture demonstrate a severalfold higher density of plasma membrane caveolae than the fibroblast-like synthetic SMCs that appear in the forming neointima after vascular injury and after 3–5 days of primary culture (Thyberg et al. 1997). Moreover, we observed that contractile SMCs are less prone than synthetic SMCs to accumulate cytoplasmic lipid droplets after exposure to low-density lipoprotein (LDL), and rather lay down excess lipids in myelin-like structures extra-cellularly, just outside caveolae (Thyberg et al. 1998). Recently, it was also found that a molecule internalized via caveolae (cholera toxin B-subunit) is more widely dispersed in contractile than in synthetic SMCs, reaching not only the endocytic pathway but also the ER and the cisternal stacks of the Golgi complex (Thyberg 2000). These variations in caveolar expression and dynamics are likely to contribute to the functional distinction between SMCs of different phenotypes, e.g., with regard to the ability to contract, proliferate, and secrete extracellular matrix components (Thyberg 1996).
In the present investigation, electron microscopic and cytochemical techniques were used to further explore the differences in cholesterol distribution and transport between contractile and synthetic SMCs. A major purpose was to find ultrastructural support for the idea that caveolae are involved in selective uptake and discharge of cholesterol at the cell surface. For this, enzymatically isolated SMCs were seeded on a substrate of fibronectin in serum-free medium and incubated in primary culture for various times, either without further additions or in the presence of cholesterol (as a part of LDL or bound to cyclodextrin). In some experiments the cells were also treated with the 3-hydroxy-3-methylglutaryl co-enzyme A (HMG CoA) reductase inhibitor mevinolin to suppress endogenous cholesterol synthesis (Alberts et al. 1980). As an alternative, the cultures were exposed to cyclodextrin to induce efflux of cellular cholesterol (Kilsdonk et al. 1995). To visualize membrane-bound cholesterol, glutaraldehyde-fixed cells were exposed to filipin and then processed for transmission electron microscopy. Filipin is a polyene antibiotic that penetrates cells, binds to cholesterol, and so produces 20–30-nm complexes that distort the structure of membranes and give them a scalloped appearance in thin sections (Severs 1997). The effects of the experimental treatments on the content and distribution of cholesterol in the plasma membrane were assessed quantitatively by counting the number of filipin-sterol complexes in cell surface domains with or without caveolae.
Materials and Methods
Materials
Ham's medium F12, newborn calf serum (NCS), and collagenase were obtained from Gibco BRL (Paisley, Scotland), and culture plastics from Nunc (Roskilde, Denmark). The culture medium was supplemented with 10 mM Hepes/10 mM Tes (pH 7.3), 50 μg/ml L-ascorbic acid, and 50 μg/ml gentamycin sulfate (medium F12). Bovine serum albumin (BSA), bovine plasma fibronectin, human low-density lipoprotein (LDL), water-soluble cholesterol (encapsulated into the hydrophobic cavity of methyl-β-cyclodextrin), mevinolin, mevalonic acid lactone, 2-hydroxypropyl-β-cyclodextrin, filipin complex, and dimethyl sulfoxide (DMSO) were from Sigma (St Louis, MO). To prepare culture substrates, fibronectin was diluted to 10 μg/ml in Dulbecco's PBS, pH 7.3, and adsorbed to the bottom of plastic dishes for 15–20 hr at 20C. Before seeding of cells, the dishes were rinsed twice with PBS and incubated with medium F12/0.1% BSA for 30 min to block unspecific binding sites. Mevinolin and mevalonic acid lactone were dissolved in 95% ethanol at 5 and 200 mM and added to the cultures in final concentrations of 5 or 500 μM, respectively (controls were given equivalent amounts of ethanol). Cyclodextrin was dissolved in culture medium at a concentration of 3% directly before use and sterile-filtered (0.2 μm).
Cell Culture
SMCs were isolated from the aortic media of 350–400-g male Sprague-Dawley rats by digestion with 0.1% collagenase in medium F12/0.1% BSA (Thyberg et al. 1990). After rinsing, the cells were seeded on a substrate of fibronectin in medium F12/0.1% BSA (40,000 cells/cm2) and incubated at 37C in a humid atmosphere of 5% CO2 in air (medium changed daily). Further details of the experimental protocols are given in connection with the description of the results. All experiments were done two to four times with similar setup and results.
Electron Microscopy
The cells were fixed in 3% glutaraldehyde in 0.1 M sodium cacodylate-HCl buffer (pH 7.3) with 0.05 M sucrose for at least 2 hr, followed by rinsing in buffer for 1–2 hr. They were subsequently exposed to 0.2 mg/ml filipin in 0.1 M cacodylate buffer (pH 7.3) for 15 hr at 20C (Orci et al. 1981). To prepare this solution, filipin was first dissolved in a small volume of DMSO and then diluted with buffer (final concentration of DMSO 1%). After renewed rinsing, the cells were scraped off the dishes with a plastic spatula, transferred to small plastic tubes, and pelleted by centrifugation. They were postfixed in 1.5% osmium tetroxide in 0.1 M cacodylate buffer (pH 7.3) with 0.7% potassium ferrocyanate for 2 hr at 4C, dehydrated in ethanol (70, 95, 100%), stained with 2% uranyl acetate in ethanol, and embedded in Spurr low viscosity epoxy resin. Sections of uniform thickness were cut with diamond knives on an LKB Ultrotome IV, picked up on formvar-coated grids, stained with alkaline lead citrate, and examined in a Philips CM120TWIN electron microscope at 80 kV (Philips; Eindhoven, The Netherlands).
Quantitation of Filipin-Sterol Complexes
To evaluate the effects of the experimental treatments on the cholesterol content of the plasma membrane, approximate mid-sagittal sections through the central parts of the cells (adjacent to the nucleus) were photographed at a final magnification of ×50,000–70,000. Within the samples so obtained, the number of caveolae on the cell surface was counted under an illuminated magnifier and the length of the plasma membrane within the plane of section was measured using a digital instrument (Calculated Industries; Yorba Linda, CA). In addition, the number of filipin-sterol complexes was counted separately in plasma membrane domains with and without caveolae.
EM of SMCs grown for 2 (
Results
Changes in Filipin Staining During Phenotypic Modification of SMCs in Primary Culture
In agreement with earlier descriptions, the newly isolated rat aortic SMCs retained a morphology characteristic of a differentiated contractile phenotype for at least 1–2 days of culture on a substrate of fibronectin in serum-free medium. During this time, a large part of the cytoplasm was occupied by myofilaments and secretory organelles, such as the ER and Golgi, were small in size (Figure 1A). Thereafter, a marked reorganization of the cells took place, with a decrease in the fractional volume of myofilaments and a corresponding growth in size of the ER and Golgi. In parallel, there was also a distinct reduction in the number of plasma membrane caveolae (Figure 1B). In the majority of the SMCs, this transition from a contractile to a synthetic phenotype was completed after 4–5 days of culture (see Hedin and Thyberg 1987; Thyberg et al. 1997; Thyberg 2000). As elaborated in detail in a separate study, treatment with 5 μM mevinolin had a strong inhibitory effect on the shift in differentiated properties of the cells. After 5 days of culture in the presence of the drug, most cells maintained a contractile phenotype as judged morphologically. Addition of 500 μM mevalonic acid lactone to the medium neutralized the effect of mevinolin and the rebuilding of the cells was completed to a similar extent as in the drug-free controls (unpublished observations).
The reactivity with filipin varied dependent on both the phenotypic state of the cells and the experimental treatments applied here. On the whole, filipin-sterol complexes were found on the cell surface, in vesicles of endosomal or lysosomal nature, in the ER (including the nuclear envelope), and in Golgi-associated vesicles. In contrast, no or only weak staining was detected in stacked Golgi cisternae and mitochondria. Both sides of the membranes were positive, but the outer part of the plasma membrane and the luminal part of the endosome/lysosome and the ER membranes showed the most prominent labeling. In contractile SMCs, the reactivity of the plasma membrane was concentrated in but not restricted to the regions where caveolae opened on the cell surface (see Thy-berg 2000). These 50–80-nm diameter invaginations were usually clustered in groups of 5–10 or more (within the plane of section), separated by regions with no or only a few such structures. Cytoplasmic vesicles and ER-like cisternae (with or without ribosomes) were frequently observed in close contact with the cell surface, particularly in regions containing caveolae. Many filipin-sterol complexes appeared at these contact sites between the plasma membrane and intracellular membrane compartments. They were present both on the extracellular/luminal and the cytoplasmic sides of the membranes, but no direct membrane continuity was noted (Figures 2A-2C). In synthetic SMCs, caveolae were fewer and filipin-sterol complexes more randomly distributed over the plasma membrane. In the cytoplasm, the strongest labeling was found in endosomes and lysosomes and portions of the ER. In the Golgi complex, stacked cisternae showed only a weak reaction and most of the staining was restricted to small and medium-sized vesicles (Figure 3). Quantitative estimations revealed that the number of caveolae per unit length of plasma membrane was 80–90% lower in synthetic than in contractile SMCs (without special treatments). However, the density of filipin-sterol complexes in caveolar membranes was the same in the two cell types. In contrast, non-caveolar membranes displayed a higher content of such complexes in synthetic than in contractile cells (Tables 1 and 2; p<0.05). No complexes of the type described above were seen in cells not treated with filipin during the EM preparation (see Figures 1A and 1B).
(
EM and filipin staining of an SMC grown for 2 days in primary culture on a substrate of fibronectin in medium F12/0.1% BSA. The cell has converted into a synthetic phenotype and a portion of the Golgi complex (G) located close to the cell surface is shown here. Filipin-sterol complexes are mainly found on the plasma membrane and in Golgi-associated vesicles/vacuoles (arrowheads). F, subplasmalemmal network of actin filaments. Bar = 200 nm.
Effects of Experimental Treatments on Filipin Staining of SMCs in Primary Culture
The cells were subjected to various experimental treatments to either increase or decrease their cholesterol content. After incubation with 100 μg/ml LDL for 2–10 hr, the number of caveoalae increased about threefold in synthetic cells but reached only about one third of the level maintained in contractile cells (Table 1). At the same time, the number of filipin-sterol complexes was augmented both on the plasma membrane and in intracellular compartments. In synthetic cells, the surface labeling was enhanced both in caveolar and non-caveolar regions, whereas contractile cells showed a clear elevation only in the latter regions (Table 1). Especially in contractile SMCs, strong staining was also observed in ER cisternae and signs of a close spatial relation between portions of the ER and the cell surface were abundant (Figure 4A). In synthetic cells, there was a further increase in the number of endosomes/lysosomes and filipin-sterol complexes were found in both the membrane and the lumen of these organelles, and in the adjacent cytoplasm (Figure 4B). A few cells in an intermediate phenotype displayed many myelin-like deposits (without evident filipin labeling) associated with the Golgi complex (Figure 5A). These structures were 50–100 nm in diameter and located either free in the cytoplasm or within the lumen of Golgi vesicles or cisternae (Figure 5B). A modest increase in the number of cytoplasmic lipid droplets was also noted, mainly in synthetic cells. Exposure of the cultures to a lower concentration of LDL (10–20 μg/ml) or cholesterol-loaded β-cyclodextrin (2 μg/ml) throughout 5 days of culture gave rise to changes similar to those just described (not shown).
Treatment of the cells with 5 μM mevinolin for 5 days partially inhibited the shift from a contractile to a synthetic phenotype and induced apoptosis in about 20% of the cells, as judged by nuclear condensation and disintegration of cytoplasmic organelles (unpublished observations). Caveolae were reduced in number in the phenotypically modified cells, but not as much as normally seen during the shift into a synthetic state (i.e., in the absence of the drug). The filipin-sterol complexes were also fewer (Table 2) and the plasma membrane had a smoother contour than in the controls (Figures 6A and 6B). Likewise, no or weak staining with filipin was noted in endosomes/lysosomes, ER cisternae, and Golgi-associated vesicles. In spite of this, a close spatial relation between caveolae and ER-like cisternae was still possible to detect. Simultaneous treatment with 500 μM mevalonic acid lactone counteracted the effects of mevinolin on the structural rebuilding of the cells as well as the reaction with filipin (Table 2). On the other hand, addition of LDL (20 μg/ml) or cholesterol-loaded β-cyclodextrin (2 μg/ml) to the culture medium did not eliminate the inhibitory effect of mevinolin on the change in cell phenotype (unpublished observations). The number of filipin-sterol complexes on the cell surface (Table 2) and in compartments such as endosomes/lysosomes and the ER was, however, equal to or even higher than in control cells kept in normal medium (Figures 7A and 7B).
Effects of LDL on the numbers of caveolae and filipin-sterol complexes in contractile and synthetic SMCs in primary culture a
SMCs were grown for 2 days in primary culture on a substrate of fibronectin in medium F12/0.1% BSA, exposed to 100 μg/ml LDL in the same medium for an additional 10 hr, and processed for EM with filipin staining. One section from each group was scanned systematically and approximate mid-sagittal sections through contractile (predominance of myofilaments and small ER/Golgi) and synthetic (predominance of ER/Golgi and few myofilaments) cells were photographed (randomly selected). The numbers of caveolae and filipin-sterol complexes in the plasma membrane were counted as described in Materials and Methods. Results are given as means of five cells in each group with SD in parenthesis (one of four experiments with similar setup and results). The statistical significance of the differences of the means was analyzed using t-test (LDL vs control).
Effects of mevinolin on the numbers of caveolae and filipin-sterol complexes in contractile and synthetic SMCs in primary culture a
SMCs were grown for 5 days in primary culture on a substrate of fibronectin in medium F12/0.1% BSA containing 5 μM mevinolin plus 0.5 mM mevalonic acid, 2 μg/ml cholesterol bound to β-cyclodextrin or 20 μg/ml LDL, and then processed for EM with filipin staining. One section from each group was scanned systematically and approximate mid-sagittal sections through contractile (predominance of myofilaments and small ER/Golgi) and synthetic (predominance of ER/Golgi and few myofilaments) cells were photographed. Numbers of caveolae and filipin-sterol complexes in the plasma membrane were counted as described in Methods. Results are given as means of four or five cells in each group with SD in parenthesis (one of four experiments with similar setup and results). The statistical significance of the differences of the means was analyzed using t-test (mevinolin vs control and mevinolin plus additions vs mevinolin).
Exposure of the SMCs to 3% β-cyclodextrin (without cholesterol) for 6 hr caused an almost total loss of caveolae and no or only very few filipin-sterol complexes were observed (Figures 8A and 8B). Otherwise, the morphology of the cells was surprisingly well preserved, but an increase in the number of apoptotic cells was noted. If the treatment with β-cyclodextrin was restricted to 2 hr, the effect on caveolae and filipin staining was less complete but still highly significant (Table 3). In contrast, no signs of cell damage were evident after this time.
Discussion
Cholesterol is an essential component in the membranes of mammalian cells and an intricate system has been developed to regulate the cellular levels of this lipid. In part, cholesterol reaches the cell via the circulation, packaged in LDL particles. These are ingested by receptor-mediated endocytosis and transferred to lysosomes, where free cholesterol is generated by enzymatic digestion. After transport out of the lysosomes, the cholesterol molecules are used for membrane biogenesis in the ER/Golgi complex and subsequently carried to the cell surface. If supplied in excess, they may also be converted into cholesteryl esters and stored in cytoplasmic lipid droplets. Cholesterol is further synthesized de novo by most cells, using enzymes residing in the ER (Brown and Goldstein 1986). To maintain membrane cholesterol at a constant level, there exists a feedback system that senses the level of cholesterol in cell membranes and, by generation of proteolytic fragments of an ER membrane protein (sterol regulatory element-binding protein or SREBP), modulates the transcription of genes encoding enzymes of cholesterol synthesis and receptors engaged in the uptake of plasma lipoproteins (Brown and Goldstein 1999).
During the past several years it has also become apparent that cholesterol is unevenly distributed in the membranes of cells, being most abundant in microdomains referred to as detergent-resistant membranes or rafts, and in cell surface caveolae (Simons and Ikonen 1997; Anderson 1998; Smart et al. 1999; Somerharju et al. 1999; Brown and London 2000). Experimental data also indicate that caveolin, a major protein constituent of caveolae, is involved in intracellular cholesterol trafficking (Fielding and Fielding 1997; Ikonen 1997; Chang and Freeman 1998). Accordingly, caveolin was shown to bind cholesterol directly (Murata et al. 1995) and to take part in its transfer from the interior to the surface of cells, either as part of vesicular carriers or caveolin-chaperone complexes moving cholesterol through the cytoplasm to caveolae (Fielding and Fielding 1996; Smart et al. 1996; Uittenbogaard et al. 1998). Other results have pointed to a role for caveolae in high- and low-density lipoprotein-mediated selective uptake and efflux of cholesterol (Fielding and Fielding 1995; Babitt et al. 1997; Ji et al. 1997; Gu et al. 2000).
EM and filipin staining of SMCs grown for 2 days in primary culture on a substrate of fibronectin in medium F12/0.1% BSA and exposed to 100 μg/ml LDL for 10 hr. (
EM and filipin staining of an SMC grown for 2 days in primary culture on a substrate of fibronectin in medium F12/0.1% BSA and exposed to 100 μg/ml LDL for 10 hr. The cell has a large Golgi complex with cisternal stacks (G) gathered in the juxtanuclear region. Lipid deposits with a myelin-like configuration are found both in the free cytoplasm (small arrow) and within the lumen of Golgi cisternae/vesicles (large arrow). In the overview (
(
EM and filipin staining of SMCs grown for 5 days in primary culture on a substrate of fibronectin in medium F12/0.1% BSA containing 5 μM mevinolin and either 20 μg/ml LDL (
(
Effects of β-cyclodextrin on the numbers of caveolae and filipin-sterol complexes in contractile and synthetic SMCs in primary culture a
SMCs were grown for 2 days in primary culture on a substrate of fibronectin in medium F12/0.1% BSA, exposed to 3% β-cyclodextrin (β-CD) for 2 hr, and then processed for electron microscopy with filipin staining. One section from each group was scanned systematically and approximate mid-sagittal sections through contractile (predominance of myofilaments and small ER/Golgi) and synthetic (predominance of ER/Golgi and few myofilaments) cells were photographed. The numbers of cell surface caveolae and filipin-sterol complexes in the plasma membrane were counted as described in Materials and Methods. Results are given as means of four or five cells in each group with SD in parenthesis (one of two experiments with similar setup and results). The statistical significance of the differences of the means was analyzed using t-test (β-CD vs control).
The findings of the present electron microscopic and cytochemical investigation using filipin as a probe indicate that cholesterol is preferentially associated with caveolae on the surface of contractile SMCs and is more randomly dispersed over the surface of synthetic SMCs. Among the cytoplasmic organelles, the most distinct labeling was found in ER cisternae (including the nuclear envelope), Golgi-associated vesicles, and endosomes/lysosomes. In contrast, no or only few filipin-sterol complexes were found in stacked Golgi cisternae and mitochondria. With regard to the Golgi complex, this is in agreement with recent observations by fluorescence microscopy showing a dot-like perinuclear pattern in SMCs exposed to filipin, distinct from the reticulate pattern seen after immunostaining for mannosidase II, a marker enzyme of medial Golgi cisternae (Thyberg 2000). These findings suggest that cholesterol passes through the Golgi complex of the SMCs via vesicular carriers rather than via stacked cisternae.
In the literature, partly different views have been expressed concerning the role of the Golgi complex in cholesterol transport. First, cholesterol synthesized in the ER has been found to reach the cell surface also when protein secretion is inhibited, implying a path that bypasses the Golgi complex (Urbani and Simoni 1990; Liscum and Munn 1999). Caveolin appears to be implicated in this process, either as a part of transport vesicles or macromolecular complexes that carry cholesterol through the cytoplasm (Smart et al. 1996; Uittenbogaard et al. 1998). This does not exclude the possibility that a smaller fraction of the newly produced cholesterol may be routed via the Golgi complex to the cell surface (Heino et al. 2000). However it is not clear to what extent the stacked cisternae are involved in this activity. Second, there is evidence for a role of trans-Golgi cisternae and/or the trans-Golgi network in handling of cholesterol after uptake of LDL by receptor-mediated endocytosis and generation of free cholesterol in lysosomes (Blanchette-Mackie and Pentchev 1998; Liscum and Munn 1999). In this case, the trans-Golgi elements operate as part of a sorting machinery involved in bidirectional transport of membrane and molecules between the cytoplasm and the cell surface (Mellman 1996).
Especially in the contractile SMCs, a close spatial relationship was noted between caveolae and ER-like cisternae or cytoplasmic vesicles. In these regions, filipin-sterol complexes appeared both on the extracellular/luminal and the cytoplasmic sides of the membranes, and sometimes also in the intervening free cytoplasm. Interestingly, such a link between caveolar and intracellular membrane compartments was observed not only in control cells but also in cells exposed to LDL or water-soluble cholesterol (increased number of filipin-sterol complexes) and in cells deprived of cholesterol by treatment with the HMG CoA reductase inhibitor mevinolin or β-cyclodextrin (reduced number of filipin-sterol complexes). It is proposed that the above-mentioned contacts serve in the translocation of cholesterol and other lipid molecules between the plasma membrane and intracellular membranes. Such a process could make it possible to pass newly synthesized cholesterol from the ER to the plasma membrane without the involvement of the Golgi complex and without the need for vesicle budding and fusion. After selective transfer of cholesterol from extracellular lipoproteins to the plasma membrane, it could likewise represent a means for movement of cholesterol from the cell surface to the ER and other cytoplasmic membrane systems.
Schematic model of non-vesicular, bidirectional cholesterol transport between the plasma membrane (caveolae) and intracellular membranes (ER) in vascular SMCs, and in particular those with a differentiated contractile phenotype. For further explanations see text.
Summing up, the findings of this and other recent studies from our laboratory indicate that the number of caveolae and the distribution/handling of cholesterol in vascular SMCs differ in a phenotype-dependent manner (Thyberg et al. 1997,1998; Thyberg 2000). SMCs in a differentiated contractile state have severalfold more numerous caveolae than cells in a dedifferentiated synthetic state. The former cells also display more frequent signs of direct or indirect (via the free cytoplasm) exchange of cholesterol and/or other lipids between the plasma membrane and the ER (without vesicle budding and fusion). Such a mechanism for cholesterol movement may function as an alternative to vesicular carriers in a cell in which most of the cytoplasm is occupied by myofilaments and the pathways for exocytosis and endocytosis therefore have a restricted capacity. A schematic model of how this lipid transport system may work is presented in Figure 9. The model is based in part on the present electron microscopic findings and in part on literature data indicating that (a) caveolin binds cholesterol directly (Murata et al. 1995), (b) caveolin-chaperone complexes are able to transport cholesterol through the cytoplasm (Uittenbogaard et al. 1998), (c) the murine scavenger receptor class B, Type I (SR-BI) is localized in caveolae and mediates exchange of cholesterol (selective uptake or efflux) between the plasma membrane and high- or low-density lipoproteins bound to the receptor (Babitt et al. 1997; Ji et al. 1997; Gu et al. 2000), and (d) cholesterol-containing rafts diffuse as small entities in cell membranes (Pralle et al. 2000).
Footnotes
Acknowledgements
Supported by the Swedish Medical Research Council, the Swedish Heart Lung Foundation, the King Gustaf V 80th Birthday Fund, and the Karolinska Institutet.
The author thanks Karin Blomgren and Birgitta Björkroth for expert technical assistance.