Human white/murine ABC8 mRNA levels are highly induced in lipid-loaded macrophages. A transcriptional role for specific oxysterols.

To identify genes that are transcriptionally activated when human macrophages accumulate excess lipids, we employed the mRNA differential display technique using RNA isolated from human monocyte-macrophages incubated in the absence or presence of acetylated low density lipoprotein and sterols (cholesterol and 25-hydroxycholesterol). These studies identified a mRNA whose levels were highly induced in lipid-loaded macrophages. The mRNA encoded the human White protein, a member of the ATP-binding cassette (ABC) transporter superfamily of proteins. The mRNA levels of ABC8, the murine homolog of the human white gene, were also induced when a murine macrophage cell line, RAW264.7, was incubated with acetylated low density lipoprotein and sterols. Additional studies demonstrated that white/ABC8 mRNA levels were induced by specific oxysterols that included 25-, 20(S)-, and 22(R)-hydroxycholesterol, and by a retinoid X receptor-specific ligand. Furthermore, the oxysterol-mediated induction of ABC8 expression in mouse peritoneal macrophages was dependent on the presence of the nuclear oxysterol receptors, liver X receptors (LXRs). Macrophages derived from mice lacking both LXRalpha and LXRbeta failed to up-regulate the expression of ABC8 following incubation with 22(R)-hydroxycholesterol. Oxysterol-dependent induction of white/ABC8 mRNA was blocked by actinomycin D but not by cycloheximide treatment of cells. We conclude that the white and ABC8 genes are primary response genes that are transcriptionally activated by specific oxysterols and that this induction is mediated by the LXR subfamily of nuclear hormone receptors. These data strongly support the hypothesis that white/ABC8 has a role in cellular sterol homeostasis.

An early event in the development of the fatty streak in the artery wall involves the entry of circulating monocytes into the subendothelial space and their subsequent differentiation into macrophages (1,2). In the presence of oxidized or modified forms of LDL, 1 these macrophages accumulate cytoplasmic lipid droplets that contain excess cholesteryl esters (1)(2)(3). The latter cells are termed macrophage "foam" cells, because the lipid droplets give them a characteristic foamy appearance when viewed under the microscope (1). The importance of monocyte/macrophages in the development of fatty streaks and the more advanced atherosclerotic lesions can be gauged from the effect of deletions or mutations of genes that are involved in either monocyte recruitment into the artery wall or their subsequent differentiation into macrophages. For example, mice defective in monocyte chemoattractant protein-1 (4), the monocyte chemoattractant protein-1 receptor (5), or macrophagecolony-stimulating factor (6, 7) exhibit reduced levels of atherosclerotic lesions when compared with normal mice.
Relatively little is known about the alterations in gene expression that occur when macrophages accumulate excess lipids to become macrophage foam cells. In previous attempts to identify such genes, macrophages were incubated with Ac-LDL to produce foam cells (8,9). These studies demonstrated that apoE (8), interleukin 8, and monocyte chemoattractant protein-1 (9) mRNAs were transiently induced in lipid-loaded foam cells. In other studies, Remaley et al. (10) demonstrated that mRNAs corresponding to Mac-2 and osteopontin, and a limited number of uncharacterized genes were induced in resident rabbit liver macrophages following the administration of a high cholesterol diet to the animals for 28 days. However, the mechanism(s) involved in these induction processes is unknown. In other studies, macrophages that were lipid-loaded in the presence of inhibitors of acyl-CoA cholesterol acyltransferase, so as to prevent the synthesis of cholesteryl esters, were shown to both accumulate unesterified cholesterol and to activate phosphatidylcholine biosynthesis (11).
More recently, specific oxysterols have been shown to be ligand activators for LXR␣ and LXR␤, two members of the nuclear hormone receptor family (12)(13)(14)(15)(16). At the present time, the only target gene that is known to be directly activated by oxysterols and LXR is cholesterol 7␣-hydroxylase, which encodes the rate-limiting enzyme in the bile acid biosynthesis pathway (17).
In contrast to the paucity of information available on the identification of lipid-or sterol-induced genes and the mechanisms involved in this process, the molecular mechanism by which excess cellular sterols repress gene expression is well established (18 -20). Excess cellular sterols result in a decrease in the maturation and nuclear localization of sterol regulatory element-binding protein and, consequently, a decrease in the transcription of sterol regulatory element-binding protein-responsive genes (18,19). Such sterol regulatory element-binding protein-responsive genes include those involved in cholesterol, fatty acid, triglyceride, and phospholipid biosynthesis (18,19).
In the current study, we have utilized the mRNA differential display technique to identify mRNAs that are induced when human macrophages are converted to macrophage foam cells. This technique identified a mRNA that encoded the human homolog of the Drosophila white gene. The human white mRNA was induced over 100-fold when macrophages were incubated with Ac-LDL, oxidized LDL, or specific oxysterols. The human white (21,22), the mouse ABC8 (23), and the Drosophila white (24) genes encode homologous proteins that are members of the ATP-binding cassette (ABC) transporter superfamily of proteins (25)(26)(27)(28). These proteins, with at least 50 members identified in humans and other eukaryotes and many other members in prokaryotes, are involved in transmembrane transport of a large variety of substrates that include peptides, phospholipids, vitamins, ions, steroid hormones, and many drugs (26 -29). Full transporter ABC proteins contain two symmetrical halves, each half containing six transmembrane domains and one ABC, whereas half transporters contain six transmembrane domains and a single ABC (26,28). Each ABC is composed of conserved Walker A and Walker B domains that are spaced by 90 -120 amino acids and a "signature" motif just upstream of the Walker B domain (28,29). These three conserved domains are required for the ABC protein to bind and hydrolyze ATP, a process that provides energy for transmembrane transport. The transmembrane domains provide substrate specificity (28). The majority of ABC transporter proteins are full transporters; they include the mammalian MDR, cystic fibrosis transmembrane conductance regulator, ABC1, and the yeast sterile 6 (STE6) protein (26,28,29). Half transporter members of this large superfamily include mammalian White/ABC8, adrenoleukodystrophy protein, peroxisomal membrane protein 70, adrenoleukodystrophy proteinrelated protein, transporter of antigenic peptides 1, transporter of antigenic peptides 2, RING4, RING11, ABC protein expressed in placenta/breast cancer resistance protein/mitoxantrone resistance protein 1, Drosophila White, Brown, Scarlet, and ABC transporter protein expressed in trachea and yeast ADP1, Pxa1p, and Pxa2p (26, 28 -40). The current demonstration that mammalian white/ABC8 mRNA levels are highly induced when macrophages are either lipid-loaded or incubated with specific oxysterols implies that this transporter protein may play an important role in macrophage sterol homeostasis.

EXPERIMENTAL PROCEDURES
Materials-DNA modification and restriction enzymes were obtained from Life Technologies, Inc. 32 P-Labeled nucleotide triphosphates were obtained from Amersham Pharmacia Biotech. Lipoprotein-deficient fetal bovine serum (LPDS) was purchased from PerImmune. The human white cDNA was a generous gift from Dr. Stylianos E. Antonarakis, University of Geneva Medical School, Switzerland. Lipopolysaccharide was purchased from Liszt, and recombinant human tumor necrosis factor ␣ was purchased from R and D Systems. Cholesterol was from steraloids and oxysterols from Sigma. The sources of all other reagents and plasmids have been given (41,42).
Mouse ABC8 cDNA-Total RNA was isolated from mouse spleen, lung, and brain tissues. The RNA was DNaseI-treated and reversetranscribed, and the cDNA was used as a template for PCR (48) using primers, described in the previous section, that are specific for the murine ABC8 cDNA (23). The 905-bp fragment, corresponding to nucleotides 1927-2831 of the ABC8 cDNA (23), was subcloned into the pCR2.1-TOPO cloning vector (Invitrogen). DNA sequence analysis indicated 100% identity with the published sequence (23) (GenBank TM accession number Z48745).
Lipoproteins-Human LDL was isolated by differential centrifugation as described (49). LDL was acetylated as described (3). Highly oxidized LDL was obtained following treatment of LDL with CuSO 4 (50). The oxidized LDL (Ͼ20 thiobarbituric acid-reactive substances/mg of protein) was dialyzed extensively, stored at 4°C, and used within 10 days.
Cell Culture-Human monocytes were isolated from peripheral blood by elutriation and plated on 100-mm dishes at a density of 1 ϫ 10 6 cells/ml in Iscove's modified Dulbecco's medium in the presence of 30% autologous serum, 0.22% insulin, antibiotics, and fungizone (49). The medium was changed on the third and sixth day. On day 8, the medium was replaced with Iscove's modified Dulbecco's medium supplemented with 10% LPDS and mevalonic acid (100 M) (medium A) in the presence of either mevinolin (5 M) or Ac-LDL (100 g/ml) and sterols (10 g/ml cholesterol (25 M) and 1 g/ml 25-hydroxycholesterol (2.5 M)). Cells incubated under these two conditions are referred to as unloaded or lipid-loaded, respectively. Where indicated, cells were incubated in medium A supplemented with either Ac-LDL, highly oxidized LDL, native LDL, specific oxysterols, cholesterol, LPS, or TNF␣. The time of incubation (4 -48 h) is indicated in the text or in the figure legend. Sterols were dissolved in ethanol and added to the cells at Ͻ1 l/ml medium. An equivalent amount of ethanol was added to control dishes. RAW264.7 cells were obtained from American Type Culture Collection and grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Cells were washed twice with phosphatebuffered saline and incubated in fresh Dulbecco's modified Eagle's medium supplemented with 10% LPDS and mevalonic acid (100 M) (medium B) and either mevinolin (unloaded cells) or Ac-LDL and/or sterols (lipid-loaded cells), as indicated above for human macrophages and as stated in legends to Figures 5 and 7. LXR␣ and LXR␤ knockout mice were generated by traditional gene targeting procedures as described (17). 2 Genotyping of tail DNA was performed by Southern blot analysis and/or PCR to confirm LXR receptor status of mice used in these studies. Mice were injected intraperitoneally with 1 ml of 4% thioglycolate solution three days prior to harvesting macrophages. To isolate macrophages, mice were sacrificed and ice-cold phosphate-buffered saline (10 ml) was injected into the peritoneal cavity. This fluid was carefully withdrawn and spun down, and the cell pellet was resuspended in high glucose Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and penicillin/streptomycin. Cells were pooled from eight mice of each genotype and distributed on duplicate plates for each treatment, and the macrophages were allowed to adhere for at least 6 h. The medium was then replaced with Dulbecco's modified Eagle's medium supplemented with 10% LPDS and mevalonic acid (100 M) Ϯ compactin (5 M) or 22(R)hydroxycholesterol (10 M) or 20(S)-hydroxycholesterol (10 M), and the cells were incubated for 42 h. Two independent experiments were performed that yielded the same results.
mRNA Differential Display and DNA Sequencing-Total RNA was isolated from unloaded or lipid-loaded human macrophages that had been incubated for 48 h in the absence or presence of Ac-LDL and sterols, as described above. The RNA was DNaseI-treated and reverse transcribed, and the PCR performed in the presence of [ 33 P]dATP, the 3Ј-primer T 11 A, and different 5Ј-primers (AP1-AP24), as described (41) using an RNAimage kit (GenHunter Corp). The PCR products were separated on a 6% denaturing polyacrylamide gel and exposed to x-ray film for 20 h. Differentially displayed bands were eluted from the gel, reamplified by PCR, and subcloned into the pCR2.1-TOPO cloning vector (Invitrogen) (41). Inserts were sequenced by the Sanger chaintermination method using the T7 Sequenase v2.0 kit (Amersham Pharmacia Biotech).
Northern Blot Assay-Total RNA was isolated using Trizol reagent (Life Technologies, Inc.), as described (41). Ten g of RNA was separated by 1% agarose/formaldehyde gel electrophoresis and transferred to a nylon membrane, and the latter was hybridized with [␣-32 P]dCTPradiolabeled DNA probes (41). Transcript abundance was determined using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA) standardized against GAPDH or cyclophilin, as indicated. For Fig. 6, the values were adjusted to define the wild-type, compactin-treated sample as 1 unit.
Southern Blot Assay-DNA was transferred to a nylon membrane, and the membrane was hybridized with [␣-32 P]dCTP-radiolabeled murine LXR␣-or LXR␤-specific probes.

Identification of a mRNA That Is Induced in Lipid-loaded
Macrophages-We hypothesized that the conversion of macrophages to macrophage foam cells in the artery wall would be associated with increased expression of specific genes that would play an important role in the development of both fatty streaks and atherosclerotic lesions. To identify such genes, we isolated human monocytes from peripheral blood and allowed them to adhere to plastic dishes and differentiate into macrophages in the presence of medium containing autologous serum (49). After one week, the cells were washed free of autologous serum and cultured for 8 -48 h in medium supplemented with 10% LPDS and mevalonic acid in the absence (unloaded) or presence (lipid-loaded) of Ac-LDL and sterols, as described under "Experimental Procedures." Fig. 1 shows that the cholesteryl ester content of the cells increased dramatically during this time, consistent with the accumulation of cytoplasmic lipid droplets and the formation of foam cells (data not shown). Total RNA was isolated from macrophages that had been unloaded or lipid-loaded for 48 h and used for mRNA differential display, as described under "Experimental Procedures." This latter technique identified a band of ϳ630 bp that was expressed in lipid-loaded but not in unloaded macrophages ( Fig. 2A). The ϳ630-bp band was identified whenever RNA from lipid-loaded cells was reverse transcribed with the T 11 A primer, and the subsequent PCR contained the T 11 A primer and a variety of 5Ј-primers. Surprisingly, differential expres-sion of the ϳ630-bp band was also observed when the 5Ј-primer was omitted from the PCR (data not shown). Based on these results, we predicted that the differentially regulated mRNA would contain a T n A sequence ϳ630 bp 5Ј of the polyadenylation site, which would be preceded by a T (i.e. TA n ) (n indicates multiple Ts or As, respectively) (Fig. 2B). The ϳ630-bp band was eluted from the gel, reamplified, and subcloned. Limited sequence analysis (172 bp) revealed Ͼ98% identity with the published sequence corresponding to the terminal 3Ј-untranslated region of the human white cDNA (21). As predicted, the human white cDNA contains a T 16 A sequence (nucleotides 2288 -2304) that is 606 bp upstream of the poly(A) site (TA 21 , nucleotides 2909 -2930) (Fig. 2B) (21). PCR performed with a T 11 A primer would thus be expected to produce a 628-bp DNA fragment.
The 628-bp cloned insert was radiolabeled and used as a probe in a Northern blot assay using RNA isolated from unloaded and lipid-loaded human macrophages. As shown in Fig.  3, the probe hybridized to a 2.9-kilobase mRNA that was present only in lipid-loaded macrophages. The data in Fig. 3 show that white mRNA levels increased during the 8 -48-h incubation of human macrophages with exogenous sterols and lipoproteins. The fold induction was difficult to determine by Northern blot assay, because white mRNA was undetectable in unloaded macrophages, even after prolonged exposure of the membrane to x-ray film. Data from semiquantitative reverse transcriptase-PCR studies indicated that white mRNA levels increased over 100-fold when macrophages were incubated for 48 h with Ac-LDL and sterols (data not shown). The membrane shown in Fig. 3 was stripped free of probe and rehybridized with radiolabeled DNA corresponding to the full-length human white cDNA (21). The results were identical to those shown in FIG. 2. mRNA differential display identifies a mRNA that is highly induced in lipid-loaded human macrophages. Duplicate dishes of human macrophages were unloaded (Ϫ) or lipid-loaded (ϩ) for 48 h, as described in the legend to Fig. 1. Total RNA was isolated from duplicate dishes (lanes 1 and 2 or 3 and 4) treated with DNaseI and used for mRNA differential display using [ 33 P]dATP, the 3Ј-primer T 11 A, and the 5Ј-primer AP1, as described under "Experimental Procedures." The PCR products were separated on a 6% denaturing polyacrylamide gel (41), and the gel was exposed to x-ray film for 20 h. A shows part of the autoradiogram. The arrow indicates a band that was observed only when the RNA was derived from lipid-loaded macrophages but not when the RNA was derived from unloaded macrophages. B shows a schematic of the human white cDNA. Nucleotides corresponding to the start and end of the open reading frame (ORF), the T 16 A at positions 2288 -2304 within the 3Ј-untranslated region, and the polyadenylation site are shown. The numbering is based on the report of Chen et al. (21). The heavy solid line corresponds to the 628-bp DNA fragment that is identified by the arrow in A. This band was isolated, subcloned, and sequenced, as described under "Experimental Procedures." Fig. 3 (data not shown). We conclude that the band identified by mRNA differential display corresponds to the human white mRNA and that this mRNA is highly induced in lipid-loaded macrophages.
Induction of White/ABC8 mRNA Requires Specific Lipids-To determine whether the induction of white mRNA was dependent on specific exogenous lipids, human macrophages were incubated for 48 h in the absence of lipids (Fig. 4, A and  B, lane 1), in the presence of Ac-LDL, cholesterol, and 25hydroxycholesterol (Fig. 4, A and B, lane 2), or in the presence of Ac-LDL, highly oxidized LDL, native LDL, specific oxysterols, cholesterol, LPS, or TNF␣ (Fig. 4). Fig. 4A shows that white mRNA levels increased when the cells were incubated in the presence of either Ac-LDL, cholesterol, and 25-hydroxycholesterol (lane 2 versus 1), 25-hydroxycholesterol (lane 3), Ac-LDL (lane 4), or highly oxidized-LDL (lane 5) but not in response to native LDL, cholesterol, LPS, or TNF␣ (lanes 6 -9 versus 1). LPS and TNF␣ are known to activate nuclear factor B and a number of signaling pathways that are responsive to increased cellular oxidative processes (53). Consistent with this role, treatment of human macrophages with LPS or TNF␣ induced the level of heme oxygenase mRNA (data not shown). However, because white mRNA levels were not induced by LPS or TNF␣ treatment (Fig. 4A, lanes 8 and 9 versus 1), we conclude that induction of white mRNA by exogenous lipids is independent of the oxidative state of the cells. Fig. 4B shows that the induction of white mRNA by 25hydroxycholesterol was dependent on the concentration of the oxysterol (compare lanes 1, 4, and 5) and that cholesterol enhanced the stimulatory effect of low levels of 25-hydroxycholesterol (Fig. 4B, lane 6 versus 4) even though cholesterol per se was inactive (Fig. 4A, lane 7 versus 1). Other oxysterols, that included 20(S)-and 22(R)-hydroxycholesterol (Fig. 4B, lanes 7  and 8), were also effective stimulants. In contrast, 22(S)-hydroxycholesterol was relatively ineffective in inducing white mRNA levels (Fig. 4B, lane 9).
In other studies, human monocytic leukemic THP-1 cells were treated with phorbol esters to promote their differentiation into macrophages (9). Lipid-loading of these macrophages with specific oxysterols and/or Ac-LDL for 48 h resulted in the induction of the levels of white mRNA, in a manner that paralleled the results shown in Figs. 3-5 (data not shown). In contrast, white mRNA was not detected in the liver-derived hepatoma HepG2 cells even after incubation of the cells for 48 h with various oxysterols (data not shown). Previous studies demonstrated that white/ABC8 mRNA was virtually absent in liver and heart but was present in a limited number of tissues and in J774 and P388D1 macrophage cell lines (21)(22)(23). Taken together, the current studies demonstrate that induction of FIG. 3. Induction of white mRNA levels in lipid-loaded human macrophages. Human macrophages were unloaded (Ϫ) or lipid-loaded (ϩ) for the indicated time period, exactly as described in the legend to Fig. 1. Total RNA was isolated, separated on a 1% agarose/formaldehyde gel, and transferred to a nylon membrane (41). The blot was hybridized with a radiolabeled probe generated from the cloned band shown in Fig. 2A. The 2.9-kilobase human white mRNA that hybridized to the probe is indicated. The blot was reprobed with radiolabeled DNA corresponding to GAPDH to ensure equivalent loading of RNA in the different lanes.  3-6), as indicated. Total RNA was isolated, and the Northern blot assay was performed as described under "Experimental Procedures," using radiolabeled DNA probes corresponding to murine ABC8 cDNA and GAPDH. macrophage white/ABC8 mRNA levels is dependent on the exogenously added oxysterol; 25-hydroxycholesterol, 20(S)-hydroxycholesterol, and 22(R)-hydroxycholesterol are more potent than 22(S)-hydroxycholesterol, and cholesterol is inactive.
The Nuclear Receptors LXR␣ and LXR␤ Are Essential for Oxysterol-mediated Induction of the White/ABC8 gene-A number of oxysterols have recently been shown to bind and activate the nuclear receptors LXR␣ and LXR␤ (13)(14)(15)(16). The relative potency of different oxysterols as activators of the LXR:RXR heterodimer is similar to their rank order in inducing the white/ABC8 mRNA levels in human and mouse macrophages (Figs. 4B and 5). These results implicate LXR␣ or LXR␤ as having a potential role in the oxysterol-dependent activation of white/ABC8.
To definitively investigate this hypothesis, peritoneal macrophages isolated from wild-type, LXR␣Ϫ/Ϫ, LXR␤Ϫ/Ϫ, or LXR␣/␤ double knockout mice were treated for 42 h with compactin ϩ mevalonic acid (unloaded), 22(R)-hydroxycholesterol ϩ mevalonic acid (sterol-loaded using an LXR agonist), or mevalonic acid alone (Fig. 6). Macrophages from wild-type or single LXR knockout mice expressed increased levels of ABC8 mRNA when treated with 22(R)-hydroxycholesterol. However, macrophages devoid of both LXR␣ and LXR␤ failed to increase the expression of ABC8 in the presence of this oxysterol. Similar results were observed when 20(S)-hydroxycholesterol was used as the sterol-loading agent (data not shown). These results confirm that LXRs play an essential role in oxysterol-mediated up-regulation of ABC8 gene expression. Furthermore, the results observed in single LXR knockout mice suggest that at least in macrophages, the two LXR genes are functionally redundant to one another, because loss of either gene alone still gives a phenotype resembling the wild-type.
Transcriptional activation of the LXR:RXR heterodimer is known to occur in response to ligands for either LXR or RXR (15,54,55). Therefore, further studies to confirm the role of this heterodimer in ABC8 expression were performed using an RXR-specific ligand LG10153. RAW264.7 cells were treated for 48 h with LG10153, 20(S)-hydroxycholesterol, or both compounds together (Fig. 7). The data show that ABC8 mRNA levels were induced when the cells were treated with either LG10153 or 20(S)-hydroxycholesterol and that the effect of both compounds on ABC8 mRNA levels was additive (Fig. 7). Similar results were obtained when these experiments were repeated with human macrophages (data not shown). These results implicate the nuclear receptor RXR as having a role in the induction of white/ABC8 mRNA levels.
Induction of White mRNA by Oxysterols Is a Primary Response-To determine whether protein synthesis is required for the sterol-dependent induction of white mRNA, human macrophages were unloaded or lipid-loaded for 8 h in the absence or presence of the protein synthesis inhibitor, cycloheximide. Northern blot assays demonstrated that the sterol-dependent induction of white mRNA was independent of protein synthesis (Fig. 8). Indeed, treatment of the cells with cycloheximide, at a dose that inhibited protein synthesis by Ͼ95%, resulted in superinduction of ABC8 mRNA levels (Fig. 8, compare lane 2 with 1 and 4 with 3). Consequently, we conclude that the induction of white mRNA by oxysterols is a primary response, because it occurs in the absence of protein synthesis.
Oxysterol-dependent Induction of ABC8 mRNA Levels Is Attenuated by Inhibitors of Transcription-Actinomycin D, a potent inhibitor of RNA polymerase II, was used to determine whether induction of ABC8 mRNA levels by oxysterols can occur in the absence of transcription. The data in Fig. 9A demonstrate that incubation of RAW264.7 cells with 20(S)hydroxycholesterol for 4 or 8 h resulted in induction of ABC8 mRNA levels (lanes 6 and 4 versus 2) by a process that was inhibited by actinomycin D (compare lane 4 with 5 and 6 with 7).
To determine whether oxysterol treatment affects the halflife of ABC8 mRNA, the experiment shown in Fig. 9B was performed. RAW264.7 cells were preincubated for 18 h in the presence of 10% LPDS to repress the ABC8 mRNA levels. Incubation was continued for an additional 6 h in the same medium supplemented with either compactin and mevalonic acid or 20(S)-hydroxycholesterol and mevalonic acid. The latter treatment induces ABC8 mRNA levels (Fig. 9B, lane 2 versus  1). Actinomycin D was then added to all dishes (Fig. 9B, 0 h), and total RNA was isolated as indicated. The relative ABC8 mRNA levels, shown in Fig. 9B, were quantitated and plotted in Fig. 9C. The results demonstrate that the half-life of the ABC8 mRNA is long (Ͼ12 h) in both sterol-deprived and oxysterol-treated cells (Fig. 9, B and C). Consequently, oxysteroldependent induction of ABC8 mRNA cannot be a result of increased stability of the ABC8 mRNA. Based on these results and on the observation that white/ABC8 mRNA levels are induced in response to a ligand for the nuclear receptor RXR (a ligand-activated transcription factor), we conclude that induc-  4). After 48 h, total RNA was isolated, and the Northern blot assay was performed using radiolabeled DNA probes specific for murine ABC8 and GAPDH, as described in the legend to Fig. 5. tion of white/ABC8 mRNA levels is dependent on increased transcription of the gene.
Murine Macrophages Express LXR␣ and LXR␤ mRNAs-The expression of LXR␣ is reportedly restricted with highest levels in the liver and lower but significant levels in the intestine, kidney, spleen, adrenals, pituitary, and adipose tissue (12,44,55). In contrast, LXR␤ is expressed in many tissues (12,44,55). To determine whether murine RAW264.7 macrophages express LXR␣ and/or LXR␤ mRNAs, total RNA was isolated from these cells treated with DNaseI and used in a reverse transcriptase-PCR, using LXR isoform-specific prim-ers. The PCR products were analyzed both by ethidium bromide staining (Fig. 10A) and by Southern blot analysis (Fig. 10, B and C). The results of Fig. 10 indicate that murine RAW264.7 macrophages express both LXR␣ and LXR␤ mRNAs. This conclusion is based on the observation that reverse transcriptase-PCRs generated bands of the expected size (Fig. 10A, 965 bp with the LXR␣-specific primers and 557 bp with the LXR␤specific primers, respectively) and that these bands hybridized to murine LXR isoform-specific DNA probes (Fig. 10, B and C). Unloaded and lipid-loaded RAW264.7 cells had similar levels of LXR␣ or LXR␤ mRNAs (data not shown).

DISCUSSION
The current studies have resulted in the identification of a mRNA that is highly induced when human or murine macrophages are incubated in the presence of Ac-LDL, oxidized LDL, or specific oxysterols. The induced mRNA encodes the human White protein or its murine homolog, ABC8. Mammalian White and ABC8 proteins are termed half transporter members of the ABC transporter superfamily of proteins, because they contain six putative transmembrane domains and one ATP-binding cassette (26 -29). White and ABC8 proteins are highly homologous to the Drosophila White protein (21)(22)(23). The latter protein dimerizes with either the Drosophila Brown or Scarlet protein to form White-Brown or White-Scarlet functional heterodimers that transport eye pigment precursors, guanine or tryptophan, respectively, into the pigment granules of the fly eye (24,56). At this time, mammalian homologs of the Drosophila brown and scarlet genes have not been identified and the  1 and 2) or lipid-loaded (lanes 3 and 4) for 8 h, as described in the legend to Fig. 1 in the absence or presence of cycloheximide (10 g/ml), as indicated. RNA isolation and the Northern blot assay were performed as described in the legend to Fig. 4.  7) and, where indicated, actinomycin D (5 g/ml). Actinomycin D was added to the medium 20 min before time 0 h, and the incubation continued for 4 or 8 h. Total RNA was isolated, and the Northern blot assay was performed as described in the legend to Fig. 5. B, RAW264.7 cells were preincubated for 18 h in medium supplemented with 10% LPDS. Incubation was continued for 6 h in the same medium supplemented with either compactin (5 M) and mevalonic acid (100 M) (lanes 1, 3, 5, 7, and 9) or 20(S)-hydroxycholesterol (10 M) and mevalonic acid (100 M) (lanes 2, 4, 6, 8, and 10). Actinomycin D (5 g/ml) was added to all dishes, and RNA was isolated immediately (0 h) or after the indicated time period. The ABC8 and GAPDH mRNA levels are shown on the Northern blot (B). C, ABC8 mRNA levels, shown in B, were normalized for loading differences and the results plotted for unloaded (E) and sterol-loaded (q) cells. There was a 6-fold increase in the level of ABC8 mRNA at the time actinomycin D was added (time 0).

FIG. 10. Murine macrophages express LXR␣ and LXR␤ mRNAs.
Total RNA was isolated from 48-h lipid-loaded RAW264.7 murine macrophages, as described in the legend to Fig. 5 and treated with DNaseI. The RNA (2.5 g) was reversed transcribed, and a fraction was used in a PCR reaction, using oligonucleotides that were specific for LXR␣ or LXR␤, as described under "Experimental Procedures." The PCR products were separated on agarose gels and either stained with ethidium bromide (A) or transferred to a nylon membrane, and the membrane was hybridized with radiolabeled DNA probes corresponding to full-length murine LXR␣ (B) or LXR␤ (C) cDNA. In A, the PCR products generated using primers specific for LXR␣ (lane 1) and LXR␤ (lane 2) are shown. S indicates the standard DNA ladder. The positions of the PCR products corresponding to LXR␣ (965 bp) and LXR␤ (557 bp) are indicated by an arrow and arrowhead, respectively. half transporter ABC protein that presumably dimerizes with White/ABC8 is unknown. However, we cannot exclude the possibility that White or ABC8 forms a homodimeric transporter. The function of mammalian White/ABC8 is also unknown. However, based on the current studies, it seems likely that macrophage White/ABC8 is involved in sterol transport and/or sterol homeostasis.
The results of studies that utilized actinomycin D (Fig. 9) or cycloheximide (Fig. 8) support the proposal that oxysterols increase the transcription of the white and ABC8 genes and that this is a primary response. The observations that white/ ABC8 mRNA levels are induced both by an RXR-specific ligand and by oxysterols that are known to activate LXR are consistent with a role for nuclear receptors in this induction process. The demonstration that RAW264.7 murine macrophages express both LXR␣ and LXR␤ mRNAs (Fig. 10) provides additional support for this latter proposal. Finally, the demonstration that up-regulation of ABC8 expression fails to occur in macrophages devoid of LXR␣ and LXR␤ (Fig. 6) definitively supports the notion that these oxysterol-activated nuclear receptors are responsible for this effect. Studies to define the ABC8 promoter elements recognized by LXR:RXR are currently underway.
Further studies will also be needed to determine whether Ac-LDL and highly oxidized LDL contain bioactive oxysterols that activate the white/ABC8 gene or whether the lipoprotein sterols are oxidized intracellularly following endocytosis of the lipoproteins into macrophages. In a recent report, Langmann et al. (57) demonstrated that the mRNA levels of ABC1, a full transporter protein, increased when human macrophages were incubated with Ac-LDL. Preliminary studies indicate that oxysterols also induce ABC1 mRNA levels. 3,4 Further studies will be necessary to determine if oxysterol-dependent induction of White/ABC8, ABC1, and perhaps other ABC proteins occurs via a common mechanism involving LXRs.
Metherall and colleagues (58,59) have proposed that MDR P-glycoprotein, an ABC transporter, has a role in cellular cholesterol homeostasis. This was based on the observation that cholesterol biosynthesis, cholesterol esterification, and the activity of the MDR P-glycoprotein were all inhibited by progesterone (58,59). This proposal is supported by the recent observation that overexpression of MDR1 in mammalian cells enhances esterification of plasma membrane-derived cholesterol (60). However, additional studies will be required to define the role of this ABC protein in cholesterol homeostasis. In other studies, treatment of cells with a number of amphiphiles, that included progesterone, led to the accumulation of cholesterol in lysosomes (61). The role of MDR, or other ABC proteins, in this process is unknown. Taken together, these observations indicate that cellular sterol homeostasis may be controlled by a number of ABC proteins that include MDR, ABC1, and White/ABC8.
The current demonstration that the levels of the white/ABC8 mRNA are highly induced in lipid-loaded or oxysterol-treated macrophages is consistent with the presence of a novel activation pathway that is likely to be important in macrophage biology. The future identification of the substrate(s) that are transported across membranes by White/ABC8 and the intracellular localization of the protein might provide important insights into the biological role of this orphan mammalian ABC transporter.