Human ABCA7 Supports Apolipoprotein-mediated Release of Cellular Cholesterol and Phospholipid to Generate High Density Lipoprotein*

Apolipoprotein-mediated release of cellular cholesterol and phospholipids was induced in HEK293 cells by expressing human ATP-binding cassette transporter A7 (ABCA7) and ABC transporter A1 (ABCA1) proteins, whether transient or stable, to generate cholesterol-rich high density lipoprotein (HDL). Green fluorescent protein (GFP) attached at their C termini did not influence the lipid release reactions. Transfected ABCA7-GFP induced apolipoprotein-mediated assembly of cholesterol-containing HDL also in L929 cells, which otherwise generate only cholesterol-deficient HDL with their endogenous ABCA1. Time-dependent release of cholesterol and phospholipid by apolipoprotein A (apoA)-I was parallel both with ABCA1 and with ABCA7 when highly expressed in HEK293 cells, but dose-dependent profiles of lipid release on apoA-I and apoA-II were somewhat different between ABCA1 and ABCA7. Analyses of the stable clones with ABCA1-GFP (293/2c) and ABCA7-GFP (293/6c) by using the same vector indicated some differences in regulation of their activities by protein kinase modulators. Dibutyryl cyclic AMP increased ABCA1-GFP and the release of cholesterol and phospholipid in 293/2c but increased neither ABCA7-GFP nor the lipid release in 293/6c. Expression of ABCA1-GFP- and apoA-I-mediated lipid release were enhanced in parallel by phorbol 12-myristate 13-acetate (PMA) in 293/2c cells. In contrast, the same treatment of 293/6c increased ABCA7-GFP, but apoA-I-mediated lipid release was significantly suppressed. Despite these different responses to PMA, all of the effects of PMA were reversed by a specific protein kinase C inhibitor Gö6976, suggesting that the changes were in fact due to protein kinase C activation. A thiol protease inhibitor, N-acetyl-Leu-Leu-norleucinal, increased the protein levels of ABCA1-GFP in 293/2c and ABCA7-GFP in 293/6c, indicating their common degradation pathway. The data indicated that human ABCA7 would compensate the function of ABCA1 for release of cell cholesterol in a certain condition(s), but post-transcriptional regulation of their activity is different.

erties and functions of specific domains. Cellular cholesterol can be derived by de novo synthesis or externally supplied via the uptake of cholesterol-containing lipoprotein particles. In contrast, most of the cells are unable to catabolize cholesterol so that cholesterol molecules must be removed from the cells and transported to the liver for their conversion to bile acids as a major exit route of the body cholesterol. Thus cholesterol transport from the peripheral cells to the liver is an essential part of cholesterol homeostasis, both for cells and for the body. High density lipoprotein (HDL) 1 is believed to play a central role in this pathway.
Lipid-free apolipoproteins with amphiphilic ␣-helical segments were demonstrated to remove cellular lipids to generate cholesterol-containing HDL (1,2). This reaction was shown deficient in fibroblasts from patients with familial HDL deficiency, Tangier disease (3,4), and therefore found essential for generation of plasma HDL. Mutations were identified in one of the members of ATP-binding cassette (ABC) transporter superfamily, ABC transporter A1 (ABCA1), as the cause of Tangier disease and other genetic HDL deficiencies (5)(6)(7)(8)(9) so that a role of this protein in the generation of HDL by apolipoprotein-cell interaction became the focus of the HDL research. ABC transporter G1 (ABCG1), an ABC transporter protein of a "half-size" structure family, has also been reported to regulate the apolipoprotein A (apoA)-I-mediated lipid release from lipid-laden macrophages (10). However, it is unclear whether this protein can generate HDL in the absence of ABCA1.
ABC transporter A7 (ABCA7) is another member of the same ABCA subfamily of "full-size" transporter as ABCA1, and its cDNA has been cloned from the human macrophage and spleen cDNA libraries exhibiting high homology to other human ABC transporters (11). ABCA7 has also been identified as the autoantigen SS-N, an epitope of Sjögren's syndrome, which was found homologous to the putative first extracellular domain of ABCA1 (12). ABCA7 mRNA and protein were induced in differentiated macrophages from human peripheral monocytes, and it was apparently expressed inversely to the cellular cholesterol level (11). Although an exact role of ABCA7 in cellular cholesterol homeostasis is unknown, it may play a relevant role in regulation of cholesterol turnover in some specific cells such as macrophages.
Recently, it was reported that mouse ABCA7 promotes apoA-I-mediated phospholipid release but failed to release cholesterol (13). Here we report that human ABCA7 mediates apolipoprotein-dependent generation of HDL by releasing both cellular cholesterol and phospholipid even in the absence of ABCA1. Although the reaction is largely similar to that of human ABCA1, apparent affinity of apolipoproteins for the cells is different. Regulations of its expression and activity by protein kinase activators are also different from those of ABCA1.

EXPERIMENTAL PROCEDURES
DNA Construction and Transfection-Full-length cDNAs for human ABCA1 and ABCA7 were cloned as described previously (12). They were introduced to pcDNA3.1/Hygro (Invitrogen), pCMV6c, and pEGFP-N1 (Clontech) to obtain constructs for the proteins without or with green fluorescent protein (GFP) at their C terminus. All the vectors have an immediate early promoter of cytomegalovirus promoter for expression of cDNA. HEK293 and L929 cells were obtained from Health Science Research Resources Bank and maintained in 1:1 mixture of Dulbecco's modified Eagle's medium and Ham's F12 medium (DF) supplemented with 10% (v/v) of fetal calf serum (FCS, Invitrogen) under a humidified atmosphere of 5% CO 2 , 95% air at 37°C. cDNAs were transfected with LipofectAMINE PLUS reagent (Invitrogen) according to the manufacturer's instruction. Cells to which the cDNAs were permanently introduced were selected with G418, and clones with high level expression of the fusion proteins were further selected according to the fluorescent intensity with GFP.
Detection of Expressed Proteins-Western blotting was carried out with an anti-ABCA1 antiserum (14,15), a rat polyclonal anti-ABCA7 antibody, and an anti-GFP antibody (Santa Cruz Biotechnology) to indicate the expression levels of the proteins. To generate the anti-ABCA7 antibody, the putative first extracellular domain of human ABCA7, amino acids 45-459, was expressed as a fusion protein with His tag at the C terminus in Escherichia coli, purified by Ni ϩ chromatography (Qiagen), 2 and used for immunization. Expression levels of ABCA1-GFP protein and ABCA7-GFP protein were also measured in situ with an FL600 fluorescent plate reader (Bio-Tek Inc.). Cellular fluorescence was measured as fluorescent intensity per unit area. GFPderived derived fluorescence was calculated by subtracting background. Intracellular localization of GFP-containing protein was examined by their fluorescence images obtained by using an Axiovert microscope (Carl Zeiss) equipped with a MicroRadiance confocal laser scanning microscope (Bio-Rad).
Apolipoproteins-ApoA-I and apoA-II were isolated from human plasma HDL fraction (density 1.09 -1.21) and stored at Ϫ80°C until use as described previously (16,17). Stock solutions (1 mg/ml) were prepared and stored at 4°C as described previously (18).
Cellular Lipid Release Assay-Cells were subcultured in 6-well trays (TPP catalogue number 92406) at a density of 1.0 ϫ 10 6 cells/well with 10% FCS-DF medium. After a 48-h incubation, the cells were washed once with buffer H (Hank's balanced salt solution containing 20 mM HEPES-KOH (pH 7.5) and 14 mM glucose), and incubated in 1 ml/well of DF medium containing 0.02% bovine serum albumin (BSA) and lipid acceptors (apoA-I, apoA-II, and 2-hydroxypropyl-␤-cyclodextrin). Buffer H supplemented with amino acids (Invitrogen catalogue number 11140-50) was used instead of DF medium in some experiments to reduce the endogenous fluorescence background in DF medium. Lipid content in the medium and cells was determined after the indicated incubation times. Procedures for lipid extraction and enzymatic assays for cholesterol and choline-phospholipids were described previously (18). Enzyme assay in combination with lysophospholipase, glycerophosphorylcholine phosphodiesterase, and choline oxidase (19) was also applied to evaluate the level of lysophosphatidylcholine (lysoPC), a causal molecule for background in choline-phospholipid assay.
Density Gradient Analysis-Cells were subcultured in 100-mm dishes (TPP catalogue number 93100) at a density of 6.0 ϫ 10 6 cells/ dish, cultured as above and stimulated with 5 ml/dish of DF medium containing 0.1% BSA and 10 g/ml apoA-I for 24 h. Medium from two dishes was combined and centrifuged to remove cell debris, and 8 ml of the supernatant was processed for sucrose density gradient ultracentrifuge (1). The solution was collected from the bottom into 13 fractions. The contents of cholesterol and choline-phospholipids as well as the density were determined for each fraction.
Statistical Analysis-Data were analyzed by one-way analysis of variance followed by Scheffé's test. A p value less than 0.05 was accepted as statistically significant.

RESULTS
Parent HEK293 cells did not respond to apolipoproteins to release either cholesterol or phospholipid (Table I). Treatment with dibutyryl cyclic AMP (dBcAMP) and phorbol 12-myristate 13-acetate (PMA) with or prior to apoA-I stimulation had no effect either (Table I and data not shown). Transient expression of ABCA1 cDNA and ABCA7 cDNA in HEK293 cells resulted in apoA-I-mediated release of both cholesterol and phospholipid in a dose-dependent manner (Fig. 1), and attachment of GFP to the C terminus of ABCA7 or ABCA1 did not influence these lipid releases (see below). For further investigation of the functions of ABCA7, we therefore obtained stable clones expressing high levels of ABCA1-GFP protein and ABCA7-GFP protein.
Release of cholesterol and phospholipid by incubation with apoA-I was demonstrated at least in three independent clones highly expressing ABCA1-GFP and also three clones expressing ABCA7-GFP, indicating that the reaction is not clonespecific but cDNA-specific. Western blotting data of parent HEK293 cells, the cells with or without transient expression of ABCA1 and ABCA7, and ABCA1-GFP-expressing clone (293/ 2c) and ABCA7-GFP-expressing clone (293/6c) are shown in Fig. 2. Neither ABCA1 nor ABCA7 protein was detected in parent HEK293 cells. ABCA1-and ABCA7-containing bands were detected at the position consistent with those of GFPattached molecules in 293/2c and 293/6c cells, as ABCA1-GFP and ABCA7-GFP at 260 -270 and 240 -260 kDa, respectively (  2F) were localized mainly in plasma membrane, as was recently reported in CHO cells stably expressing rat ABCA7 (21) and in HEK293 cells transiently expressing mouse ABCA7 (13).
Typical profiles of lipid release by apoA-I and apoA-II from 293/2c and 293/6c cells are shown in Fig. 3. LysoPC assay 2 R. Aoki and K. Ueda, unpublished results.

TABLE I
ApoA-I-mediated lipid release from parent HEK293 cells treated with protein kinase stimulants Cells were subcultured in 6-well trays at a density of 1.0 ϫ 10 6 cell/well and incubated for 48 h. The cells were washed with buffer H, and 1 ml/well of 0.1 % BSA-DF containing apoA-I (10 g/mL), dBcAMP (300 M) and PMA (320 nM) was added as indicated. Medium was collected after 24 h for cholesterol (Ch) and phospholipid (PL) analysis (g/well). Results shown are the average and variation for two samples. confirmed that most of the choline-phospholipids released to the medium in the absence of apolipoproteins was lysoPC and that its level was independent of apolipoprotein concentration (data not shown). The dose-dependent curve of the reaction by apoA-II was similar to that of apoA-I with respect to molar concentration of the proteins for both cholesterol and cholinephospholipid (Fig. 3). These results indicated that ABCA7 directly promotes both cholesterol and phospholipid efflux to apoA-I, just as ABCA1 does. The releases of phospholipid and cholesterol appear parallel by increasing concentrations of apoA-I and apoA-II in 293/2c (ABCA1-GFP) and 293/6c (ABCA7-GFP). The lipid release seems to reach the maximum at lower concentration of apolipoprotein with the ABCA1-expresing cells (293/2c) than the ABCA7-expressing cells (293/6c). The EC 50 for the apoa-I mediated release of cholesterol and phospholipid from 293/6c cells was ϳ4.5 and 4.2 g/ml (0.16 and 0.15 M), respectively, whereas 1.8 and 1.2 g/ml (0.064 and 0.043 M) for that from 293/2c cells ( Fig. 3 and data not shown). Time course analysis of apoA-I-mediated lipid release supported the idea of simultaneous release of phospholipid and cholesterol both from 293/2c cells and from 293/6c cells as cholesterol and phospholipid in the medium were well detectable as early as 1 h after apoA-I stimulation, and they increased linearly for 8 h (Fig. 4). Density gradient analysis of the medium demonstrated that both cholesterol and phospholipid were recovered in the fractions with a density peak at around 1.08 g/ml, indicating that HDL particles were generated from the ABCA7-GFP-expressing cells similarly to the ABCA1-GFPexpressing cells (Fig. 5). Expression of ABCA7-GFP also induced cholesterol release in L929 cells. We reported elsewhere that L929 cells release phospholipid but not cholesterol by apoA-I (22). As shown in Fig. 6, parent L929 cells released only phospholipid to apoA-I even at high concentrations of apoA-I (open symbols). Stable expression of ABCA7-GFP protein in L929 cells caused substantial release of cholesterol together with enhancement of phospholipid release (closed symbols). The level of endogenous ABCA1 expression was not affected in L929 by transfection and expression of ABCA7 (Fig. 6C).
Both dBcAMP and PMA enhanced apoA-I-mediated lipid release and ABCA1-GFP protein level in 293/2c. In the presence of dBcAMP (300 M) or PMA (320 nM) with apoA-I, significant increase was induced in cholesterol and phospholipid release from 293/2c cells (Fig. 7, A and B, open columns). The increase of the lipid release was correlated with elevation of ABCA1-GFP protein level evaluated by GFP-derived fluorescence intensity and Western blotting analysis (Figs. 7C and 8A, lanes 2-7). The effects of dBcAMP and PMA were slightly synergistic, although not additive, in the condition tested.
In contrast, treatment with dBcAMP did not affect apoA-Imediated cholesterol or phospholipid release in 293/6c cells, and PMA decreased release of both lipids (Fig. 9, A and B, shadowed columns). On the other hand, apoA-I and ALLN caused the increase of ABCA7-GFP in 293/6c, although the increase in cholesterol and phospholipid release was not statistically significant (Figs. 8B, 9, A and B, hatched columns). ABCA7-GFP protein level was not significantly affected either by dBcAMP, whereas it was slightly up-regulated by PMA (Figs. 8B and 9C). None of the compounds tested influenced non-specific cholesterol release to 2-hydroxypropyl-␤-cyclodextrin from 293/2c or 293/6c (data not shown). These effects of dBcAMP, PMA, and apoA-I were similar in HEK293 cells transiently or stably expressing ABCA7 (Fig. 8B, lanes 1-4, and data not shown) Induction of ABCA1 and ABCA7 by dBcAMP or PMA was further investigated in 293/2c and 293/6c cells, respectively, by monitoring their GFP-derived fluorescence (Fig. 10). In 293/2c cells, PMA induced the transient increase of ABCA1-GFP fluorescence, as the increase was evident after 2 h, reaching a peak at 8 -12 h (Fig. 10A). dBcAMP-induced fluorescence increase was rather continuous at least up to 30 h after a time lag of 4 -6 h. The effect of these two compounds were additive. Continuous monitoring of cellular fluorescence in the fluorescence-free medium yielded similar results (data not shown). Being consistent with these results, the apoA-I-mediated lipid release from 293/2c for 4 h was not affected by dBcAMP but was enhanced by PMA (data not shown). In 293/6c cells, fluorescence level was changed neither by dBcAMP nor by PMA (Fig.  10B). Fluorescent levels were increased in parallel with lipid release in 293/2c cells when treated with ALLN, whereas the increase was slight and unsustainable in 293/6c (data not shown).
Gö6976, a PKC inhibitor (25), reversed all of the changes caused by PMA. In 293/2c cells, the enhancement of apoA-Imediated lipid release and ABCA1-GFP induction was inhibited ( Figs. 8A and 11, A-C). In 293/6c cells, suppression of the apoA-I-mediated lipid release by PMA was recovered (Fig. 11,  D and E), whereas the ABCA7-GFP level was reduced to the control level (Figs. 8B and 11F). Gö6976 alone had no signifi- cant effects on apoA-I-mediated lipid release from 293/2c or 293/6c (Fig. 11). DISCUSSION The function and its regulation of human ABCA7 was studied by using its expressing system of in HEK293 cells. ABCA7 exhibited a function for generation of cholesterol-containing HDL upon the interaction with apoA-I and apoA-II so much as human ABCA1 does. Response of ABCA7 to protein kinase modulators, dBcAMP and PMA, was somewhat different from ABCA1. ABCA1 protein level was increased by either reagent, and its function for mediating the apoA-I-mediated lipid release was increased in parallel, whereas ABCA7 and its activity were not increased by dBcAMP. Interestingly, ABCA7 was slightly increased by PMA, but its activity for mediating lipid release by apolipoprotein was rather suppressed. All of the effects of PMA were reversed by PKC-specific inhibitor Gö6976, indicating that they are mediated by PKC.
Deficiency of ABCA1 causes loss of plasma HDL, as demonstrated in the patients with Tangier disease (5-9) and in ABCA1 knockout mice (26,27), to indicate that there is no significant compensatory backup system for supply of plasma HDL. At the cellular level, however, ABCG1 may function as a  Hygro (A, lanes 1 and 2), ABCA1-GFP/pcDNA3.1 (A, lanes  3 and 4), ABCA7/pCMV6c (B, lanes 1 and 2), and ABCA7-GFP/ pcDNA3.1 (B, lanes 3 and 4) as in Fig. 1. Cells were collected after 30 h for crude membrane preparation. Membrane proteins (80 g for A, lanes 1 and 2 and B; 40 g for A, lanes 3 and 4) were separated on a 5.5% SDS-polyacrylamide gel, transferred onto Immobilon (Millpore), and analyzed by using rabbit anti-ABCA1 antiserum (A) and rat polyclonal anti-ABCA7 antibody (Ab) (B). C-E, Western blotting analysis of 293/2c, 203/6c, and parent HEK293 cells. Cells were cultured in 10% FCS-DF and processed to prepare crude membrane fraction. 20 g of membrane proteins were separated as above and analyzed by using rabbit anti-ABCA1 antiserum (C), rat polyclonal anti-ABCA7 antibody (D), and mouse monoclonal anti-GFP antibody (E). regulator of lipid transport in lipid-laden macrophages, although its in vivo function is not defined. Suppression of ABCG1 expression with the ABCG1-specific antisense oligonucleotide caused 32 and 25% reduction of the release of cholesterol and phospholipid, respectively, whereas ABCA1 expression was not down-regulated (10). It was also shown that levels of ABCG1 mRNA in non-cholesterol-laden macrophages from two patients with Tangier disease were significantly greater than controls, although the function of ABCG1 was not examined (28).
We demonstrated that human ABCA7 mediates the apolipoprotein-dependent cellular lipid release and consequent assembly of new HDL in a very similar manner as human ABCA1 in vitro. Our experimental protocol fulfills the condition to observe the effect of ABCA7 in the absence of ABCA1 so that this is an isolated function of this protein, at least in vitro. The results from L929 cells expressing ABCA7-GFP in addition to their endogenous ABCA1 (Fig. 6) suggested that the effects of ABCA1 and ABCA7 in apolipoprotein-mediated HDL generation may be synergistic. However, it is not clear whether ABCA1 and ABCA7 generate HDL particles independently, cholesterol-deficient HDLs and cholesterol-containing HDLs in this case. The reaction mechanisms of ABCA1 and ABC7 should be further investigated, including their difference, cross-talk, and physiological relevance.
In HEK293 cells, release of lipid may reach maximum at a slightly lower concentration of apoA-I and apoA-II (2-5 g/ml) when ABCA1 was expressed than in the cells expressing ABCA7 (maximum at 20 -30 g/ml apoA-I) (Figs. 1 and 3). The results with ABCA1 were apparently consistent with previous findings with RAW264 where ABCA1 was strongly induced by dBcAMP, as shown by oligonucleotide array analysis that the mRNA level was increased 10-fold, whereas ABCA7 mRNA level was low and not affected (18). A similar dose-dependent curve was observed with human fibroblast WI-38 cells in which ABCA1 was found (22) but not ABCA7 (data not shown). Further investigation is required for understanding the underlying mechanism for this apparent difference in kinetic profiles of the HDL assembly reactions between ABCA1 and ABCA7.
HEK293 cells transiently expressing mouse ABCA7 released phospholipids but not cholesterol by apoA-I, even in the presence of scavenger receptor-BI or loading of extra cholesterol mass (13). As HEK293 cells transiently expressing mouse ABCA1 were able to generate cholesterol-containing HDL (13,29), the difference may be between human ABCA7 and mouse ABCA7.
The effect of dBcAMP on ABCA1-GFP level in 293/2c is consistent with the previous reports, an increase of ABCA1 mRNA in human fibroblasts (22), RAW264 cells (18,30), and macrophages (31) by cAMP analogues. It has been reported that there is a cAMP-responsive element in the ABCA1 promoter (32). However, neither ABCA1 protein ( Fig. 1 and data not shown) nor apoA-I-mediated HDL generation (Table I) was detected in parent HEK293 cells even after the dBcAMP treatment so that it is unlikely that the enhancement of the ABCA1 activity by dBcAMP in 293/2c is carried out by this cAMPresponsive element of the endogenous ABCA1 gene. In fact, no immunoreactive band was detected in 293/2c at the position of lower molecular weight than ABCA1-GFP even after the cAMP treatment. The promoters of the transfected cDNAs were common (immediate early promoter of cytomegalovirus) for ABCA1-GFP and ABCA7-GFP and dBcAMP did not increase ABCA7-GFP so that the effect of cAMP in this case is likely to be on the post-transcriptional modulation.
ABCA1 protein expressed in Xenopus oocytes was phosphorylated by PKA in a cell-free system (33), and its activity as an anion transporter was enhanced after short term treatment with PKA activators (33). Also, 8-bromo-cAMP promoted phosphorylation of ABCA1 by a 1-h incubation in normal human fibroblasts and increased the apoA-I-mediated release of cholesterol and phospholipid without changing its mRNA or pro-tein level (34). A more recent report indicated that phosphorylation by PKA at a specific site of ABCA1 is constitutive but important for the apoA-I-mediated phospholipid release (35). Our results seem rather consistent with the cellular conditions similar to Xenopus oocytes (33) or fibroblasts (34). Regulation of ABCA7 by PKA may then be different in this regard.
The effect of the PKC activator, PMA, also differentiated the response of ABCA1 and ABCA7. ABCA1-GFP in 293/2c was increased by PMA, and the activity also seemed increased in parallel (Fig. 7). In contrast, PMA decreased the apoA-I-mediated lipid release from 293/6c expressing ABCA7-GFP (Fig. 9,  A and B), whereas it did not cause significant reduction of the GFP-derived fluorescence (Figs. 9C and 11F). These effects of PMA were all reversed by Gö6976, an inhibitor of Ca 2ϩ -dependent isoform(s) of PKC (25). The data indicate that specific activities of ABCA7 can be modulated by PKC.
Turnover of ABC transporters has not been fully understood. We recently reported that degradation of ABCA1 is protected by apolipoproteins from the degradation by thiol protease, most likely calpain (15,24). ALLN was effective to increase ABCA1 protein expressed in HEK293 cells, indicating that ABCA1 protein expressed by the exogenously transfected cDNA is metabolized by the similar mechanism, being consistent with other reports (23,24). The results with the transfected ABCA7 were largely similar to those with ABCA1, showing that the metabolic pathway of these proteins are common.
The physiological relevance of the activity of ABCA7 reported in this study should be further investigated. ABCA1 distribution is ubiquitous (33), and its dysfunction results in the loss of plasma HDL (5-9), whereas tissue distribution of ABCA7 is reportedly restricted to myelo-lymphatic tissues in human (11) and mouse (36) or preferentially in platelets in rat (21) and mouse (13). Therefore, the ABCA7-mediated lipid release may not contribute significantly to a source of plasma HDL. However, it may still play an important role in cellular cholesterol homeostasis in particular tissues including macrophages. In Tangier disease, accumulation of cholesteryl ester is found in foamy histiocytes in the reiticuloendotherial system, fibroblasts of the cornea, melanocytes, Schwann cells, neurons, and non-vascular-muscle cells (37). It is interesting to point out that Tangier disease patients appear to be only at moderately increased cardiovascular risk despite the almost complete loss of plasma HDL and considerable cholesterol ester accumulation in resident macrophages of many tissues (37). ABCA1 knockout mice showed tissue distribution of lipid deposition identical to Tangier disease and no abnormalities in aorta even in aged mice (38). These findings may suggest that there is a protecting system against the development of atherosclerosis even in the absence of ABCA1, and the ABCA7-mediated lipid  293/6c (B). The clone cells were subcultured into 12-well trays 5.0 ϫ 10 5 cells/well and cultured for 64 h in 10% FCS-DF. Reagents indicated were added into the wells at the time 16, 40, 52, 60, 62, and 63 h after subculture. The final concentrations of dBcAMP and PMA were 300 M and 320 nM, respectively. At 64 h of the incubation after subculture, medium was removed, cells were washed with buffer H, and cellular fluorescence was measured. Data are relative to control cells that were maintained in 10% FCS-DF throughout the experiment. In this experimental condition, a 48-h incubation with dBcAMP did not affect cell growth, whereas PMA suppressed it by about 15%. These effects were same for parent HEK293 cells, 293/2c cells, and 293/6c cells. Results represent means Ϯ S.D. for three samples. release from macrophages in the vascular wall is one of the candidates. In fact, ABCA7 protein is detected in peripheral blood monocytes after in vitro differentiation into macrophages followed by acetylated LDL loading (11). Specific roles of this protein should be examined, especially in atherosclerotic tissues.
Further study of the mechanism by which ABCA7 mediates the release of cholesterol and generation of HDL should provide us with important information for intracellular trafficking and homeostasis of cholesterol by comparing it with ABCA1. It would also lead us to novel strategies for treatment of atherosclerosis. Controlled induction of ABCA7 in certain specific organs, tissues, or cells, such as macrophages and lymphomyeloid cells, would be efficient to remove cholesterol from peripheral tissues to prevent atherosclerosis, by itself or in coordination with ABCA1.