ATP-binding Cassette Transporter A1 Mediates Cellular Secretion of (cid:1) -Tocopherol*

(cid:1) -Tocopherol ( (cid:1) -TOH) is associated with plasma lipoproteins and accumulates in cell membranes throughout the body, suggesting that lipoproteins play a role in transporting (cid:1) -TOH between tissues. Here we show that secretion of (cid:1) -TOH from cultured cells is mediated in part by ABCA1, an ATP-binding cassette protein that transports cellular cholesterol and phospholipids to lip-id-poor high density lipoprotein (HDL) apolipoproteins such as apoA-I. Treatment of human fibroblasts and murine RAW264 macrophages with cholesterol and/or 8-bromo-cyclic AMP, which induces ABCA1 expression, enhanced apoA-I-mediated (cid:1) -TOH efflux. ApoA-I lacked the ability to remove (cid:1) -TOH from Tangier disease fibroblasts that have a nonfunctional ABCA1. BHK cells that lack an active ABCA1 pathway markedly increased secretion of (cid:1) -TOH to apoA-I when forced to express ABCA1. ABCA1 also mediated a fraction of the (cid:1) -TOH efflux promoted by lipid-containing HDL particles, indicating that HDL promotes (cid:1) -TOH efflux by both ABCA1-dependent and -independent processes. Exposing apoA-I to ABCA1-expressing cells did not enhance its ability to remove (cid:1) -TOH from cells ABCA1, consistent with this transporter participating directly in the translocation of (cid:1) -TOH to apolipoproteins. These studies provide evidence that ABCA1 mediates secretion of cellular a process that may facilitate influence

␣-Tocopherol (␣-TOH) is associated with plasma lipoproteins and accumulates in cell membranes throughout the body, suggesting that lipoproteins play a role in transporting ␣-TOH between tissues. Here we show that secretion of ␣-TOH from cultured cells is mediated in part by ABCA1, an ATP-binding cassette protein that transports cellular cholesterol and phospholipids to lipid-poor high density lipoprotein (HDL) apolipoproteins such as apoA-I. Treatment of human fibroblasts and murine RAW264 macrophages with cholesterol and/or 8-bromo-cyclic AMP, which induces ABCA1 expression, enhanced apoA-I-mediated ␣-TOH efflux. ApoA-I lacked the ability to remove ␣-TOH from Tangier disease fibroblasts that have a nonfunctional ABCA1. BHK cells that lack an active ABCA1 pathway markedly increased secretion of ␣-TOH to apoA-I when forced to express ABCA1. ABCA1 also mediated a fraction of the ␣-TOH efflux promoted by lipid-containing HDL particles, indicating that HDL promotes ␣-TOH efflux by both ABCA1dependent and -independent processes. Exposing apoA-I to ABCA1-expressing cells did not enhance its ability to remove ␣-TOH from cells lacking ABCA1, consistent with this transporter participating directly in the translocation of ␣-TOH to apolipoproteins. These studies provide evidence that ABCA1 mediates secretion of cellular ␣-TOH into the HDL metabolic pathway, a process that may facilitate vitamin transport between tissues and influence lipid oxidation.
␣-Tocopherol (␣-TOH) 1 is the most chemically and biologically active form of vitamin E, an essential fat-soluble nutrient with antioxidant properties. Because of its hydrophobicity, ␣-TOH is co-transported with lipids in plasma lipoproteins and accumulates in cell membranes throughout the body (1). In human plasma approximately half of the ␣-TOH is located in high density lipoproteins (HDL), with the other half being present in low density lipoproteins (LDL) and very low density lipoproteins (VLDL) (1). Phospholipid transfer protein facilitates the transfer of ␣-TOH between serum lipoproteins and from artificial liposomes to HDL (2). Although it distributes among all major lipoprotein subclasses, some cells take up ␣-TOH preferentially from HDL (3)(4)(5)(6). This uptake appears to be mediated by a HDL receptor called scavenger receptor B1 (SR-B1) (4), which also selectively transports HDL cholesteryl esters into cells (7). These findings implicate HDL as a major transporter of ␣-TOH between tissues.
Processes involved in the secretion of ␣-TOH from cells are still poorly understood. ␣-TOH transfer protein (␣-TTP), a cytosolic liver protein identified as a product of the causative gene for familial isolated vitamin E deficiency (8), plays an important role in liver membrane transport and secretion of ␣-TOH (9). Arita et al. (10) showed that a hepatoma cell line expressing ␣-TTP secreted ␣-TOH more efficiently than cells lacking ␣-TTP. They also showed that brefeldin A, which inhibits VLDL secretion by disrupting the Golgi apparatus, had no effect on ␣-TOH secretion. Based on these results, the authors suggested that hepatic ␣-TTP stimulates the secretion of cellular ␣-TOH into the extracellular medium via a novel non-Golgi-mediated pathway that may be linked to cellular cholesterol metabolism and/or transport (10). To date, however, this suggested pathway has not been identified. Moreover, mechanisms by which extrahepatic cells secrete ␣-TOH are unknown.
Previous studies suggested that HDL and/or its major apolipoprotein, apolipoprotein A-I (apoA-I), plays a role in promoting secretion of ␣-TOH from cells. Panzenbock et al. (11) showed that lipid-free apoA-I promotes ␣-TOH efflux from macrophages in parallel to and just as effectively as cholesterol and phospholipid efflux. Lipid-poor apoA-I removes cellular cholesterol and phospholipids by an active transport pathway controlled by an ATP-binding cassette transporter called ABCA1 (12). Mutations in the ABCA1 gene cause Tangier disease (13)(14)(15)(16), a severe HDL deficiency syndrome characterized by accumulation of sterols in tissue macrophages and prevalent atherosclerosis (17). Mice lacking a functional ABCA1 also have a severe deficiency of fat-soluble vitamins including vitamin E (18). These studies are consistent with ABCA1 being involved in mobilizing ␣-TOH from cells into the HDL pathway.
Here we used cell lines expressing different levels of ABCA1 to test the possibility that this transporter mediates ␣-TOH secretion from cells. Results show that efflux of ␣-TOH from cells to HDL apolipoproteins is a function of ABCA1 expression, further implicating the HDL pathway in the transport of ␣-TOH between tissues.

EXPERIMENTAL PROCEDURES
Lipoproteins and ApoA-I-LDL and HDL 3 (herein referred to as HDL) were prepared by sequential ultracentrifugation in the density range of 1.019 to 1.063 and 1.125 to 1.21 g/ml, respectively, and HDL was depleted of apoE and apoB by heparin-agarose chromatography (19). ApoA-I was purified from HDL and delipidated as described previously (20). Trypsinized HDL was prepared as described previously (21) by treating HDL with trypsin for 30 min at 37°C at an HDL: trypsin protein ratio of 40:1. This procedure digests ϳ20% of the total HDL protein content of HDL without disturbing its lipid composition. LDL was acetylated by the method of Goldstein et al. (22).
Cell Culture-Human fibroblasts were obtained from skin explants from a normal subject and a homozygous Tangier disease patient and were immortalized as described previously (16,23). The Tangier disease cells expressed a nonfunctional ABCA1 because of a homozygous Arg-527 3 Trp substitution (16). Murine RAW264 macrophages and baby hamster kidney (BHK) cells were obtained from ATCC (Manassas, VA). BHK cells expressing human ABCA1 were generated using the mifepristone-inducible GeneSwitch system (Invitrogen, Carlsbad, CA). Cells were transfected initially with the pSwitch plasmid using Fugene6 (Roche Molecular Biochemicals). Clonal lines were isolated and then transfected with pGene/V5-His/lacZ and assayed for ␤-galactosidase activity with the ␤-Gal assay kit (Invitrogen). Clonal lines that gave the highest mifepristone-induced activity were then transfected with the human ABCA1 gene. The full-length open reading frame of the ABCA1 gene was subcloned into pGene/V5-HisA, and the plasmid was linearized before transfection. Control BHK cells were derived from the same clonal line but lacked the ABCA1 plasmid. All cells were grown and maintained in DMEM containing 10% fetal bovine serum until experimental treatments.
The following protocols were used to induce ABCA1 in the different cell lines. Fibroblasts were cholesterol-enriched by incubation with DMEM containing 2 mg/ml fatty acid-free bovine serum albumin (BSA) and 30 g/ml cholesterol (from a 10 mg/ml ethanol solution) for 48 h followed by 20 -24-h incubations in DMEM/BSA containing 1 mM 8-Br-cAMP (16,23). RAW264 macrophages were incubated for 24 h with DMEM/BSA plus 50 g/ml acetylated LDL followed by 20 -24 h incubations with DMEM/BSA plus 0.3 mM 8-Br-cAMP (24). ABCA1 transfected BHK cells were incubated for 24 h in DMEM/BSA plus 10 nM mifepristone.
Immunoblotting-Control and ABCA1-transfected BHK cells were solubilized in 50 mM Tris buffer containing 1% SDS, 0.1 M mercaptoethanol, and 0.5 mM EDTA, and proteins were resolved by 6% polyacrylamide gel electrophoresis. Proteins were transblotted onto nitrocellulose, and ABCA1 was identified with a C-terminal antibody using an ECL detection assay (24).

RESULTS
Previous studies showed that lipid-free apoA-I removes ␣-TOH from cultured cells (11). To test if this involves ABCA1, we compared the ability of apoA-I to remove radiolabeled ␣-TOH and cholesterol from fibroblasts expressing different amounts of ABCA1. Treatment of immortalized human fibroblasts with cholesterol plus 8-Br-cAMP markedly induces ABCA1 expression and apoA-I-mediated lipid efflux (16,23). In contrast, fibroblasts from patients with Tangier disease have a nonfunctional ABCA1 whether or not they are treated with cholesterol/8-Br-cAMP (16,23). When [ 14 C]␣-TOH-or [ 3 H]cholesterol-labeled normal and Tangier disease fibroblasts were incubated for 6 h with BSA alone, a higher fraction of ␣-TOH than cholesterol appeared in the medium (Fig. 1), indicating that ␣-TOH diffuses from cells at a faster rate than cholesterol does. Addition of apoA-I to BSA-containing medium caused only a small or negligible increase in both ␣-TOH and cholesterol efflux from untreated (control) normal fibroblasts (Fig. 1, A and C) and from both control and cholesterol/8-Br-cAMPtreated Tangier disease cells (Fig. 1, B and D). In contrast, when ABCA1 was induced in normal cells by cholesterol/8-Br-cAMP, apoA-I-mediated efflux of both ␣-TOH and cholesterol increased 5-fold. Dose-response curves revealed that this induced efflux of ␣-TOH occurred by a high affinity (K d ϳ10 Ϫ8 M) saturable process (Fig. 2), as is the case for ABCA1-mediated efflux of cholesterol and phospholipids (25). Thus, secretion of ␣-TOH from fibroblasts to apoA-I depends on an induced and functional ABCA1.
To examine further the role of ABCA1 in mediating ␣-TOH secretion, we compared the ability of different lipid acceptors to remove ␣-TOH from cholesterol/8-Br-cAMP-treated normal and Tangier disease fibroblasts. Lipid-free apolipoproteins remove cholesterol from cells by the ABCA1 pathway, whereas phospholipid-containing particles remove cholesterol by ABCA1-independent mechanisms (25)(26)(27). As in Fig. 1, addition of apoA-I to the medium increased ␣-TOH efflux from normal but not from Tangier disease fibroblasts (Fig. 3). The addition of HDL particles, which contain both phospholipids and apolipoproteins, increased ␣-TOH efflux from both normal and Tangier disease cells, but efflux was increased to a greater extent from normal cells. Trypsinized HDL, which lacks active apolipoproteins (21), also stimulated ␣-TOH efflux from cells, but there were no differences between normal and Tangier ABCA1-mediated ␣-Tocopherol Secretion disease cells. These results indicate that phospholipids in HDL particles also promote ␣-TOH efflux, but this occurs by an ABCA1-independent mechanism.
Previous studies showed that treatment of RAW264 macrophages with cAMP analogs induces ABCA1 expression and apoA-I-mediated cholesterol and phospholipid efflux (24,28). We tested for whether cAMP also stimulates apoA-I-mediated efflux of ␣-TOH from macrophages. As with control fibroblasts, a higher fraction of ␣-TOH than cholesterol diffused from untreated RAW264 macrophages in the presence of BSA alone, and the addition of apoA-I had little or no effect on efflux of either compound (Fig. 4). Treatment of these cells with 8-Br-cAMP caused a significant increase in ␣-TOH and cholesterol efflux only in the presence of apoA-I, indicating that this cAMP analog selectively induced apoA-I-mediated efflux of both compounds. These results show that apoA-I removes ␣-TOH from macrophages by the cAMP-inducible ABCA1 pathway.
We overexpressed ABCA1 in cells to determine whether this would enhance ␣-TOH secretion. For these studies, we transfected BHK cells with an inducible ABCA1 gene, radiolabeled cells with [ 14 C]␣-TOH, induced ABCA1 expression, and measured apoA-I-mediated ␣-TOH efflux. Control BHK cells contain virtually no detectable ABCA1, whereas induced cells express very high levels of this protein (Fig. 5A, inset). Based on immunoblots, this expression level is 5-10-fold higher than maximum levels in human fibroblasts or murine macrophages. ApoA-I promoted significant cholesterol and phospholipid efflux only from the induced cells (not shown). Inducing ABCA1 expression caused only a small increase in ␣-TOH efflux in the presence of BSA alone. ApoA-I removed very little ␣-TOH from control cells but markedly stimulated ␣-TOH efflux from ABCA1 transfectants (Fig. 5A), indicating that apoA-I-mediated ␣-TOH secretion requires expression of ABCA1.
Two possible mechanisms could account for the role of ABCA1 in transporting cellular ␣-TOH to lipid-free apoA-I. First, ABCA1 could transfer phospholipids to apoA-I, generating acceptors for ␣-TOH that diffuses from cells by ABCA1independent processes. Second, ABCA1 could transfer ␣-TOH directly along with other lipids to apolipoproteins. To distinguish between these mechanisms, we incubated unlabeled con- trol and ABCA1-transfected BHK cells for 4 h with apoA-I, transferred the media to [ 14 C]␣-TOH-labeled control cells (low ABCA1), and measured [ 14 C]␣-TOH efflux during the second incubations. Results showed that apoA-I had little ability to promote ␣-TOH efflux from control cells whether pre-incubated with cells expressing very low or high levels of ABCA1 (Fig.  5B). Therefore, under these incubation conditions, apoA-I did not acquire enough phospholipid mass by the ABCA1 pathway to become an acceptor for ␣-TOH that diffuses from cells. This makes it likely that ABCA1 participates directly in the transport of ␣-TOH from cells to apolipoproteins.
We also examined the possibility that ABCA1 may be involved in transporting ␣-TOH into cells. The results showed that there was no difference in the cellular uptake of [ 14 C]␣-TOH between normal and Tangier disease fibroblasts (data not shown). Moreover, none of the treatment protocols used in this study, which were inserted between the labeling and efflux incubations, had a significant effect on the cellular content of [ 14 C]␣-TOH. It appears therefore that ABCA1 mediates the unidirectional transport of ␣-TOH from cells to suitable extracellular acceptors. DISCUSSION Previous studies showed that lipid-free apoA-I promotes secretion of ␣-TOH from cells (11). Here we show that this occurs by the ABCA1 lipid secretory pathway. Three lines of evidence support this conclusion. First, treatment of human fibroblasts and murine RAW264 macrophages with cholesterol and/or 8-Br-cAMP, which induces ABCA1 expression and apoA-I-mediated lipid removal (16,23), also enhanced apoA-I mediated ␣-TOH efflux. Second, apoA-I lacked the ability to remove ␣-TOH from Tangier disease fibroblasts that have a nonfunctional ABCA1. Third, BHK cells that lack an active ABCA1 pathway in the basal state markedly increased secretion of ␣-TOH to apoA-I when forced to express ABCA1.
The current study shows that mature HDL particles are also capable of promoting ␣-TOH efflux from cells. Comparisons of normal and Tangier disease cells exposed to native and trypsintreated HDL revealed that this efflux occurs by both ABCA1dependent and -independent processes. The rate of diffusion of ␣-TOH from cells exceeds that observed for cholesterol, which may reflect the comparatively greater hydrophilicity of the vitamin. These findings indicate that lipoproteins can remove ␣-TOH from cells by both passive and active processes.
The diffusible properties of ␣-TOH raised the possibility that the effects of ABCA1 on ␣-TOH secretion were indirect. By lipidating apoA-I, ABCA1 may have generated acceptors for ␣-TOH that passively desorbs from the plasma membrane. Our pulse-chase studies, however, showed that exposing apoA-I to cells expressing high levels of ABCA1 did not increase its ability to remove ␣-TOH from cells lacking ABCA1. These findings indicate that, during the brief incubations used in these studies, ABCA1 did not lipidate apoA-I enough to generate acceptors for ␣-TOH that diffuses from cells. Thus, ABCA1 appears to be involved directly in transporting ␣-TOH. It is possible that ABCA1 co-transports ␣-TOH, cholesterol, and phospholipids to the cell surface where the lipid complex is solubilized and removed by apolipoproteins (12).
ABCA1 plays a gatekeeper role in transporting tissue cholesterol into the HDL metabolic pathway (12). The current study suggests that this protein also functions to transport ␣-TOH taken up by tissues back into plasma via apoA-I and HDL. Although ␣-TOH distributes among all major lipoprotein subclasses, HDL is the preferred vehicle for transport of ␣-TOH to at least some tissues (3)(4)(5)(6). This tissue uptake appears to utilize the SR-B1 receptor system (4), which also selectively transports cholesteryl esters from HDL into cells (7). Interest-ingly, expression of SR-B1 is suppressed when liver cells accumulate ␣-TOH, consistent with feedback repression of ␣-TOH delivery into cells by this pathway (29). We found that supplementing macrophages and fibroblasts with ␣-TOH had no effect on ABCA1 expression (data not shown), implying that ␣-TOH does not regulate its secretion by this transporter. Taken together, these studies suggest that SR-B1 and ABCA1 operate in concert to transport ␣-TOH between tissues by the HDL pathway.
ABCA1 is highly expressed in the liver, where it may play a role in secreting dietary ␣-TOH delivered to the liver from the intestine via apoB-containing lipoproteins. Hepatocytes contain ␣-TTP, which retains ␣-TOH, transports the vitamin between membranes, and promotes its secretion (9,10). Because much of the ␣-TOH secreted from hepatocytes in vivo is associated with nascent VLDL, it was assumed that ␣-TTP facilitates incorporation of ␣-TOH into VLDL during its assembly and secretion. Arita et al. (10), however, showed that ␣-TTP stimulates secretion of ␣-TOH from cultured hepatocytes by a pathway distinct from lipoprotein assembly but linked to cholesterol metabolism and/or transport. Our results suggest that this pathway involves ABCA1 and suitable extracellular acceptors of ␣-TOH. The secreted ␣-TOH may then distribute among lipoproteins by a process facilitated by phospholipid transfer protein (2).
The ABCA1 pathway may also help generate ␣-TOH in tissues to alleviate oxidative stress. Macrophages ingest large amounts of cell-derived cholesterol and other lipids within tissues with high rates of cell turnover and at sites of inflammation. In the artery wall, macrophages take up oxidized lipoproteins, leading to atherogenic foam cell formation. Cholesterol accumulation in macrophages induces ABCA1 expression (30), which would be predicted to increase secretion of ␣-TOH along with cholesterol and phospholipids. In support of this idea is a study by Asmis and Jelk (31) showing that cholesterol loading of cultured macrophages depletes them of ␣-TOH. Whether this renders cells more susceptible to subsequent oxidation is not clear at present. Although some intracellular ␣-TOH may be localized to sites for release into the extracellular environment, other pools of ␣-TOH may protect cells against oxidation. The ␣-TOH transported from cells by ABCA1 may help protect co-secreted and other extracellular lipids from oxidation. Increased secretion of cellular ␣-TOH by lipid-laden macrophages could also help explain why lipoproteins in diseased artery walls contain relatively normal amounts of vitamin E despite the fact that their lipids are oxidized (32).
The current study shows that lipophilic compounds other than cholesterol and phospholipids utilize the ABCA1 pathway for secretion from cells. The robust induction of ABCA1 by sterols implies that these molecules are primary targets for secretion by this transporter. Our results suggest that secretion of at least some cellular pools of ␣-TOH is tightly linked to sterol metabolism and transport. It remains to be determined whether this also applies to other fat-soluble vitamins and lipophilic compounds.