Synergistic Transcriptional Activation of Human Acyl-coenzyme A: Cholesterol Acyltransterase-1 Gene by Interferon-g and All-trans- Retinoic Acid THP-1 Cells*

Acyl-coenzyme A:cholesterol acyltransferase (ACAT) is an intracellular enzyme involved in cellular cholesterol homeostasis and in atherosclerotic foam cell formation. Human ACAT-1 gene contains two promoters (P1 and P7), each located in a different chromosome (1 and 7) (Li, B. L., Li, X. L., Duan, Z. J., Lee, O., Lin, S., Ma, Z. M., Chang, C. C., Yang, X. Y., Park, J. P., Mohandas, T. K., Noll, W., Chan, L., and Chang, T. Y. (1999) J. Biol Chem. 274, 11060–11071). Interferon-g (IFN-g), a cytokine that exerts many pro-atherosclerotic effects in vivo, causes up-regulation of ACAT-1 mRNA in human blood monocyte-derived macrophages and macrophage-like cells but not in other cell types. To examine the molecular nature of this observation, we identified within the ACAT-1 P1 promoter a 159-base pair core region. This region contains 4 Sp1 elements and an IFN-g activated sequence (GAS) that overlaps with the second Sp1 element. In the monocytic cell line THP-1 cell, the combination of IFN-g and all-trans-retinoic acid (a known differentiation agent) enhances the ACAT-1 P1 promoter but not the P7 promoter. Additional experiments showed that all-trans-retinoic acid causes large induction of the transcription factor STAT1, while IFN-g causes activation of STAT1 such that it binds to the GAS/Sp1 site in the ACAT-1 P1 promoter. Our work provides a molecular mechanism to account for the effect of IFN-g in causing transcriptional activation of ACAT-1 in macrophage-like cells.

ACAT 1 is an intracellular enzyme responsible for catalyzing the intracellular formation of cholesteryl esters from choles-terol and long-chain fatty acyl-coenzyme A (1). In mammals, two ACAT genes have been identified (2)(3)(4)(5). In adult human tissues, ACAT-1 is the major enzyme present in various tissues, including macrophages, liver (hepatocytes and Kupffer cells), and adrenal gland (6,7). ACAT-1 is also present in the intestine; however, the major enzyme involved in the intestinal cholesterol absorption may be ACAT-2, which is mainly located in the apical region of the intestinal villi (7). The relative tissue distributions of ACAT-1 and ACAT-2 in mice and monkeys are not entirely consistent with those found in humans (8,9) raising the possibility that the distribution of the two ACATs in various tissues may be species dependent. In macrophages and other cell types, a dynamic cholesterol-cholesteryl ester cycle exist; the formation of intracellular cholesteryl esters is catalyzed by ACAT-1, while the hydrolysis of cholesteryl esters is catalyzed by the enzyme neutral cholesteryl ester hydrolase (10,11). The net accumulation of intracellular cholesteryl esters is affected at the substrate level, as well as at the levels of the enzymes ACAT and neutral cholesteryl ester hydrolase (12)(13)(14). The main mode of sterol-specific regulation of ACAT-1 has been identified at the post-translational level, involving allosteric regulation by its substrate cholesterol (1,15). On the other hand, the cellular and molecular nature of non-sterolmediated ACAT-1 regulation remains largely unknown. Recently, using mouse macrophage-derived foam cells, Panousis and Zuckerman (12) reported that IFN-␥ increased the cellular cholesteryl ester content and reduced high density lipoproteinmediated cholesterol efflux; its cellular effects were attributed to its ability to increase ACAT-1 message (12) and to induce down-regulation of the Tangier Disease gene (the ABC1 transporter) (16). In the current work, we showed that IFN-␥ increased ACAT-1 message and protein content in human monocyte-derived macrophages. To examine the molecular mechanism of IFN-␥ action on ACAT-1 gene regulation in macrophages, we identified the important cis-acting elements in the human ACAT-1 P1 promoter. In order to perform transient transfection experiments, we used THP-1 cell, a monocytic human cell line as the cell model. Upon treatment with retinoids, including all-trans-retinoic acid (ATRA), THP-1 cells differentiate into macrophage-like cells (17)(18)(19)(20). Our results show that ATRA and IFN-␥ synergistically caused up-regulation of ACAT-1 gene expression. Additional experiments revealed that ATRA causes increased gene expression of the transcription factor STAT1, while IFN-␥ is essential to cause STAT1 to undergo phosphorylation dependent dimerization and to bind to the GAS site present in the ACAT-1 P1 promoter.

EXPERIMENTAL PROCEDURES
Cell Culture and Treatments-Human monocytes were isolated according to a published procedure (21) with slight modification: human leukocyte packs were obtained from Shanghai Blood Service Center and used within 1 day. The cells were diluted (2:1, v/v) with cold phosphatebuffered saline (PBS), layered on an equal volume of Ficoll-Paque (Amersham Pharmacia Biotech), and centrifuged for 20 min at 2,500 rpm at room temperature. Mononuclear cells were collected and washed three times at 4°C (to remove platelets) by adding 100 ml of PBS followed by centrifugation at 1,000 rpm for 10 min. The remaining red blood cells in the pellet were lysed by treatment with 10 ml of 0.2% NaCl for 45 s, followed by sequential additions of 10 ml of 1.6% NaCl and 30 ml of cold PBS. The pelleted cells were suspended in cold RPMI 1640 with 7% human type AB serum to a density of 5 ϫ 10 6 /ml, plated onto 60-mm tissue culture dishes that were precoated with 2 ml/dish of FBS, and incubated for 90 min at 37°C. Next, the dishes were washed three times with warm RPMI 1640 (37°C) to remove unadhered cells. The adhered cells were judged to be more than 95% monocytes by ␣-naphthylacetate esterase staining. The cells were cultured for up to 16 days in RPMI 1640 medium supplemented with 7% human type AB serum, with a medium change every other day. Other cell lines were from ATCC. Cells were incubated 60-mm dishes in a 37°C incubator with 5% atmospheric CO 2 . All media were supplemented with 100 g/ml kanamycin, 50 units/ml streptomycin, 2 g/liters sodium bicarbonate, plus 10% fetal bovine serum (FBS). THP-1 and U937 cells were grown in RPMI 1640 medium. HepG2 and Caco-2 cells were grown in Dulbecco's modified Eagle's medium. HEK293 cells were grown in minimal essential medium medium. Chinese hamster ovary cell lines AC29 and 25RA (22,23) were grown in F12 medium.
Transfection and Luciferase Assay-A series of ACAT-1 P1 promoter/ luciferase reporter (Luc) constructs were transfected into THP-1 or U937 cells using the DEAE-dextran method (29,30). After washing twice with PBS, 1 ϫ 10 6 cells were transfected with 1.5 g of ACAT-1 promoter/Luc plasmid and 0.75 g pCH110 as internal control in 1 ml of STBE (25 mM Tris-HCl, pH 7.4, 5 mM KCl, 0.7 mM CaCl 2 , 137 mM NaCl, 0.6 mM Na 2 HPO 4 , 0.5 mM MgCl 2 ) containing 150 g of DEAEdextran. The cells were incubated for 20 min at 37°C, washed once with RPMI 1640 without FBS, then resuspended in 5 ml of fresh RPMI 1640 with 10% FBS, and plated at 2 ϫ 10 5 cells/ml/well in a 24-well plate for 40 h. HepG2, CACO-2, and HEK293 cells were transfected by the methods of calcium phosphate co-precipitation essentially as described by Liu et al. (31). Briefly, cells were plated at 1 ϫ 10 5 cells/well in 1 ml of medium in 24-well tissue culture plates 1 day before transfection. One h before transfection, cells were replaced with fresh medium. Calcium phosphate precipitates containing (per well) 0.3 g of ACAT-1 promoter/Luc and 0.15 g of pCH110 were prepared. The DNA/calcium phosphate precipitates were incubated with the cells at 37°C for 8 h, after which time the cells were washed once with PBS, and replaced with Dulbecco's modified Eagle's medium or minimal essential medium containing 10% FBS. After incubation for 7 h, cells were treated with or without IFN-␥ (100 units/ml), or ATRA (10 Ϫ6 M), or IFN-␥ (100 units/ml) plus ATRA (10 Ϫ6 M). 40 h later, the cells were harvested and the cell pellets were lysed in 200 l of lysis buffer (Reporter lysis buffer, Promega, catalog number E397A), vortexed for 5 s, and spun at 2000 ϫ g for 5 min at room temperature. 60 l of the cell lysate was mixed with 60 l of luciferase assay buffer (Promega) for luciferase activity measurement (Promega Instruction Bulletin Part number TB101) in an Auto Lumat BG-P luminometer (MGM Instrument Inc.). For ␤-galactosidase activity assay, the luminescent ␤-galactosidase detection Kit II was used (CLONTECH User Manual PT2106-1).
RNA Preparation and Northern Blot Analysis-THP-1 cells were cultured at 2 ϫ 10 5 /ml in 60-mm dishes. Human blood monocytes were cultured at 1.5 ϫ 10 6 /60-mm dishes. Cells were treated with IFN-␥ as indicated for 40 h before harvest. The preparation of total RNA was

FIG. 1. A 159-bp core region comprises the basal transcriptional activity of ACAT-1 P1 promoter. A, Luc constructs (bars shown on the left panel) containing serial 5Ј and 3Ј deletions as indicated between
Ϫ598 and ϩ34 of the ACAT-1 P1 promoter were cotransfected with pCH110 into THP-1 cells. The cells were harvested 48 h after transfection for activity assays. The luciferase activity per each cell extract was normalized by using the ␤-galactosidase value found in the same cell extract. The reporter construct activities shown on the right panel were expressed as relative luciferase activities, using the value of the reporter activity driven by the SV40 promoter as one. Values were means of triplicate determinations. Sizes of error bars indicated 1 S.E. B, nucleotide sequence analysis of ACAT-1 P1 promoter core region. The four Sp1 elements were boxed. Asterisks indicate the three major transcriptional initiation sites (Li et al. (26)). Sequence of Exon 1 was underlined.
FIG. 2. Four Sp1 elements are functionally present in ACAT-1 P1 promoter. A, Luc constructs containing serial 5Ј-and 3Ј-deletion (bars shown on the left) of the 159-bp core region were transfected into THP-1 cells. The luciferase activities (shown on right) were determined in the same manner as described in the legend to Fig. 1A. The promoterless plasmid pGL2-E was used as a negative control. B, Luc constructs containing single or multiple Sp1 mutations (marked by the ϫ) of the 159-bp core region were transfected into THP-1 cells. The luciferase activities shown on right panel were determined as described in the legend to Fig. 1A. C, EMSAs using nuclear extracts of THP-1 cells. The wild-type and Sp1-1234-mutant DNA fragments of ACAT-1 P1 promoter (depicted at the left panel) were, respectively, labeled and 1 ϫ 10 4 dpm of labeled probe was used for each binding reaction. Lane 1, 32 P-labeled wild-type DNA as probe alone. Lane 2, binding reaction between labeled wild-type DNA probe and nuclear extracts. Lane 3, competition by adding 100-fold molar excess of cold probe to the binding reaction described for lane 2. Lane 4, competition by adding 100-fold molar excess of nonspecific DNA to the binding reaction described for lane 2. Lane 5, supershift reaction by adding 1 l of anti-Sp1 antibody to the binding reaction described for lane 2. Lanes 6 -10, the same conditions as described for lanes 1-5, except using the Sp1-1234-mutant DNA as the labeled probe.
according to the single step acid guanidinium thiocyanate phenol chloroform method (Trizol Regent, Life Technologies, Inc.). Total RNA, 20 g/ sample, were electrophoresed in a 1% agarose gel containing 2.2 M formaldehyde and transferred to a Nytran membrane (Schleicher and Schuell, Dassel, Germany) with 3.0 M sodium chloride, 0.3 M sodium citrate (20 ϫ SSC) as the transfer buffer. The membrane was cross-linked by UV irradiation and incubated for 10 min at 65°C in 0.5 M sodium phosphate buffer (pH 7.2), 7% SDS, 1 mM EDTA (prehybridization buffer). The polymerase chain reaction products of human ACAT-1 cDNA (1486 -2686, 1.2-kilobases) and human GAPDH cDNA (291-bp) were used as templates for labeling probes. Labeled probes were made with [␣-32 P]dATP by the random primer method using a Random-labeling Kit (Promega). Blots were prehybridized and hybridized with labeled probes and washed under high stringency conditions. Hybridization was carried out at 65°C in the same solution as prehybridization, except for the addition of labeled probe. The membrane was washed with 40 mM sodium phosphate buffer (pH 7.2), 0.1% SDS at room temperature for 5 min three times, and at 65°C for 20 min. After washing, the membrane was exposed and the intensity of the bands was quantified by densitometric analysis using the UVP Labwork software (UVP Inc.). To serve as control, rehybridization of the same blot with the human GAPDH probe was carried out. The sample mRNA expression levels were normalized by the intensity of the human GAPDH mRNA bands.
Western Blot Analysis-Cells were harvested with 10% SDS in 50 mM Tris, 1 mM EDTA (pH 7.5) with 25 mM dithiothreitol, and incubated at 37°C for 20 min, then sheared with a syringe fitted with an 18-gauge needle. Protein concentration of the cell extract was determined by a modified Lowry method (34). The affinity purified anti-ACAT-1 IgGs (designated as DM10) was used as the primary antibodies against ACAT-1 (35). Western blots, using freshly prepared cell extracts in SDS, were conducted according to a previously described procedure (35).
ACAT Activity Assay-The assay was performed essentially as described previously (15,35). AC29, 25RA, and THP-1 cells were cultured at 2 ϫ 10 5 /ml in 60-mm dishes and then treated in various manners for 40 h as indicated. The ACAT-1-deficient mutant cell line AC29, and its parental cell 25RA derived from Chinese hamster ovary cells (2,15,35) were used to ensure that the ACAT activity assayed in vitro work properly. For THP-1 cells, the suspended and adherent cells were collected by direct centrifugation and scrapping, respectively, at room temperature. The two groups of cells collected from the same dish were pooled together, washed with PBS once, and centrifuged to collect the cell pellets. Cold 1 mM Tris, 1 mM EDTA (pH 7.8), at 100 l/sample, was added to cell pellet chilled on ice. The mixtures were left on ice for 5 min. Brief but vigorous vortexing (30 s to 1 min) was used to cause extensive cell lysis. The protein concentration of the cell homogenates was kept at 2-4 mg/ml in buffer A (50 mM Tris, 1 mM EDTA at pH 7.8 with protease inhibitors). The enzyme was solubilized and assayed in mixed micelle condition as previously described (35).

A 159-bp Core Region with 4 Sp1 Elements Is Responsible for
Human ACAT-1 P1 Promoter Activity-Human ACAT-1 gene is located in two different chromosomes (1 and 7), each chromosome containing a separate ACAT-1 promoter (P1 and P7). Northern analyses have revealed the presence of four ACAT-1 mRNAs (7.0, 4.3, 3.6, and 2.8-knt) in all the human tissues and cell lines examined (3). The 2.8 and 3.6-knt messages are produced from the P1 promoter, while the 4.3-knt mRNA is produced from two different chromosomes by a novel RNA recombination event that presumably involves trans-splicing (26). The P1 promoter is contiguous with the coding sequence and spans from Ϫ598 to ϩ65 of the ACAT-1 genomic DNA (26).
To determine the minimal region of the P1 promoter, we trans-FIG. 3. Effect of IFN-␥ on the ACAT-1 gene expression during the human blood monocyte-macrophage differentiation process. Human blood monocytes were cultured at 1.5 ϫ 10 6 / 60-mm dish for various days as indicated and then treated with or without IFN-␥ (100 units/ml) for 40 h before harvested for preparation of total RNA and protein extract. A and B, quantitation by RT-PCR (26 to 32 cycles). Appropriate primers described under "Experimental Procedures" were used to obtain the ACAT-1 P1 promoter transcript (designated as the P1 product), the ACAT-1 P7 promoter transcript (designated as P7 product), and the transcript for the control gene (indicated as GAPDH) by RT-PCR. Control experiments indicated that between cycles 25 and 35, the ACAT-1 P1 transcript and the ACAT-1 P7 transcript could be estimated semiquantitatively by RT-PCR (data not shown). The ratio of DNA contents (shown at the bottom panels) was determined by using the UVP Labwork software (UVP Inc.). C, immunoblotting of ACAT-1 protein from extracts of blood monocyte-derived macrophages treated with or without IFN-␥ (100 units/ml) for 40 h. Protein extracts were prepared and immunoblotting were conducted as described under "Experimental Procedures." Samples used (40 g of protein/ lane) were freshly prepared with SDS. The membrane was incubated with DM10 (0.5 g/ml) as the primary antibody. The immunoreactive proteins were visualized using the ECL detection system and autoradiography. The intensities of bands were determined by using the UVP Labwork software (UVP Inc.). The data are expressed as relative ACAT-1 protein level using the value in untreated cells as 1.0. The ratios of the ACAT-1 protein from treated or untreated cells were shown at the bottom panel.
fected THP-1 cells with constructs containing various deleted fragments fused upstream to a luciferase reporter gene of the pGL2-enhancer vector (pGL2-E), and measured luciferase activities. As shown in the right panel of Fig. 1A, the results indicated that the maximal transcriptional activity is located within the 159-base pair from Ϫ125 to ϩ34 (Fig. 1A). Sequence analysis by computer revealed 4 Sp1 elements are located in this core region (Fig. 1B). We next performed various specific deletion analyses to test the relative importance of these 4 Sp1 elements. The results (Fig. 2, A and B) showed that the most important basal transcription activity is present in the first two Sp1 elements from the 5Ј-end of the 159-bp core region.
To demonstrate the functional importance of the 4 Sp1 elements involved in basal transcription, we next performed EMSA. We used nuclear extracts of THP-1 cells and the 159-bp DNA fragments containing mutations in each of all the 4 Sp1 elements as labeled probes. The bindings of labeled probes were tested by competing with unlabeled wild-type or mutated probes in 100-fold molar excess. The results (Fig. 2C) illustrated that the wild-type DNA fragment formed several DNAprotein complexes (lanes 2, 4, and 5); the bindings were eliminated upon incubation with excess unlabeled probe (lane 3) and supershifted by incubation with the anti-Sp1 antibodies (lane 5). Additional control experiments showed that the fragment containing mutations in all 4 Sp1 elements had no specific binding (lanes 7-10).

Interferon-␥ Causes up-regulation of Human ACAT-1 Expression in Blood
Monocyte-derived Macrophages-The human blood monocytes were incubated in culture for up to 16 days. This procedure causes monocytic cells to differentiate into mature macrophages within several days. Cells incubated for various time points were treated with or without IFN-␥ for 40 h. The total RNAs and proteins of treated and untreated cells were extracted for RT-PCR and Western blot. The results show that both the ACAT-1 P1 promoter transcript and the ACAT-1 protein level increased during the monocyte differentiation process; and these increases were further augmented in cells treated with IFN-␥ (Fig. 3, A and C). In contrast, the level of the human ACAT-1 P7 promoter transcript was not significantly altered throughout the time course of the experiment (Fig. 3B).
The Combination of IFN-␥ and ATRA Is Needed to Enhance ACAT-1 P1 Promoter Activity in THP-1 Cell-Using THP-1 cell, we tested the functional responses of the ACAT-1 P1 promoter toward IFN-␥, ATRA, or a combination of both. As shown in Fig. 4A, treating cells with IFN-␥ and ATRA, but not with IFN-␥ or with ATRA alone, synergistically enhanced the luciferase expression driven by the ACAT-1 P1 promoter. To investigate the cell type and promoter specificity of this effect, we tested the human ACAT-1 P1, P7, and SV40 promoters in THP-1 cells, using the luciferase reporter activity assays. The results (Fig. 4A) showed that neither IFN-␥, ATRA, nor their combination, had any detectable effect on the ACAT-1 P7 or the SV40 promoter. We also tested the potential effect of ATRA and/or IFN-␥ on ACAT-1 P1 promoter in other human cell lines including HEK293, HepG2, CaCO-2, and U937. The results (Fig. 4B) demonstrated that the synergistic effect of ATRA and IFN-␥ occurred only in the monocytic cell lines (e.g. U937, THP-1), but not in other cell types (e.g. HEK293, HepG2, or CACO-2). Using THP-1 cells, we next investigated the time and dose requirements. For IFN-␥, the results showed that the enhancement exhibited a saturable process, with maximal enhancement seen when IFN-␥ reached 500 units/ml (Fig. 5A). For ATRA, a non-saturable, linear relationship between concentration and effect was observed (Fig. 5B). The effect of IFN-␥ and ATRA continued to increase within the time frame examined (60 h, Fig. 5C).
To ascertain that the activating effect of ATRA/IFN-␥ on ACAT-1 promoter bears biological relevance, we treated THP-1 cells for 40 h with either IFN-␥ or ATRA, or a combination of both, then examined the ACAT-1 at the levels of mRNAs, protein, and enzyme activities. Semiquantitative RT-PCR analysis showed that the level of human ACAT-1 P1 promoter transcripts in treated cells increased by about 3-fold (Fig. 6A), while the level of human ACAT-1 P7 promoter transcripts were not significantly altered (Fig. 6B). Consistent with these results, Northern blotting (Fig. 6C) showed that the amounts of 3.6-and 2.8-knt mRNAs (26) were significantly enhanced by the combination treatment of IFN-␥ and ATRA. In addition, the ACAT-1 protein as analyzed by Western blotting (Fig. 7A), as well as by the ACAT enzyme activity, measured in cholesterolindependent manner, was all significantly and proportionally increased (Fig. 7B).
IFN-␥ Activated Sequence (GAS) Is Required for the Synergistic Effect by IFN-␥ and ATRA-Sequence analysis by computer showed that the core region of human ACAT-1 P1 promoter contained an GAS that overlaps exactly with the second Sp1 element from 5Ј-end (Fig. 8A). To test its functional significance, a series of P1 promoter deletion and point mutation constructs were made, linked to a luciferase reporter gene, and used in transient transfection studies in THP-1 cells. The results indicated that the two constructs containing the GAS element (at the top of Fig. 8A) responded to IFN-␥ and ATRA, while the shorter promoter lacking the GAS element and first two Sp1 elements (at the bottom of Fig. 8A) did not. Specific   FIG. 6. IFN-␥ and ATRA synergistically increase ACAT-1 mRNA. Total RNAs were prepared from THP-1 cells treated for 40 h with or without IFN-␥ (100 units/ml), or with ATRA (10 Ϫ6 M) or with IFN-␥ (100 units/ml) plus ATRA (10 Ϫ6 M), respectively. A and B, quantitation by RT-PCR (26 to 32 cycles). Appropriate primers described under "Experimental Procedures" were used to obtain the ACAT-1 P1 promoter transcript (designated as the P1 product), the ACAT-1 P7 promoter transcript (designated as P7 product), and the transcript for the control gene (indicated as GAPDH) by RT-PCR. Control experiments indicated that between cycles 25 and 35, the ACAT-1 P1 transcript and the ACAT-1 P7 transcript could be estimated semiquantitatively by RT-PCR (data not shown). The ratio of DNA contents (shown at the bottom panels) was determined using the UVP Labwork software (UVP Inc.). C, quantitation by Northern analysis. 20 g of total RNAs per lane from cells treated in various manners as indicated was employed, using a 32 P-labeled ACAT-1 cDNA probe; the same filter was rehybridized with a 32 P-labeled human GAPDH cDNA probe. After exposing with PhosphorImager, the intensities of the 2.8-and 3.6-knt ACAT-1 were normalized to that of the GAPDH mRNA levels; the intensities of bands were determined by using the UVP Labwork software (UVP Inc.). The ratios of the 2.8-and 3.6-knt message from cells treated in various manners as indicated were shown on the right panel.
mutations in the GAS element, but not mutations in the Sp1 elements, abrogated the synergistic effect by IFN-␥ and ATRA (Fig. 8B). Therefore, the GAS element, rather than the 4 Sp1 sites plays an important role in mediating the regulatory response to IFN-␥ and ATRA.
To further examine the functional significance of the GAS element, we isolated nuclear extracts from THP-1 cells cotreated with IFN-␥ and ATRA, and performed EMSA using the wild-type P1 promoter (the 159-bp DNA) as the labeled probe. As shown in Fig. 8, C and D, two specific bands, one migrating slower than the other, were detectable (lane 2). These two bands were abolished by preincubation with unlabeled competitors containing all the Sp1 and GAS elements (lane 3). As shown in Fig. 8C, gel supershift assays using either anti-Sp1 antibodies (lane 4) or anti-STAT1 antibodies (lane 5), or both antibodies (lane 6), indicated that these two bands were specific complexes formed between STAT1 and Sp1. When excess unlabeled probes containing either the first or the second Sp1 element were used as competitors, the two bands were also competed out (lanes 7 and 8). When unlabeled probe containing only the first two mutant or all four mutant Sp1 elements were used as competitors, both bands moved faster (lanes 9 and 10) than those in the control lane (lane 2). These two bands were supershifted by using the anti-Sp1 antibody (lane 11) but not by using the anti-STAT1 antibody (lane 12). Additional experiments (Fig. 8D) showed that when labeled wild-type DNA fragments were used as probe and unlabeled DNA fragment containing all four wild-type Sp1 elements and the mutant GAS (GASm) as competitors, one band was found to move faster than the control lane (comparing lanes 4 -6 with lane 2). This band was supershifted by adding anti-STAT1 antibody, but not by adding anti -Sp1 antibody (comparing lanes 4 -6). When labeled probe containing mutant GAS was used, two bands moved faster (lanes 8); they were supershifted by anti-Sp1 antibody but not by anti-STAT1 antibody (comparing lanes  10 and 11). Together, these results demonstrate that the first two Sp1 sites and the GAS site are functionally important, and that the second Sp1 site overlaps with a GAS site to form a novel overlapping GAS/Sp1 element. This GAS/Sp1 element is recognized by both STAT1 and Sp1 present in the nuclear extracts of treated THP-1 cells.
ATRA Induces STAT1 Expression, While IFN-␥ Causes the STAT1 to Dimerize and Bind to GAS Element in the ACAT-1 Promoter-STAT1 is a key component of the IFN-␥-dependent transcriptional activation complex (36,37). We examined the transcript level of STAT1 in control and treated THP-1 cells by RT-PCR. As shown in Fig. 9A, STAT1 transcript was not detectable in control THP-1 cells (lane 1). Treating cells with ATRA gave rise to a remarkable increase (lane 3), while treating cells with IFN-␥ caused only a modest increase in the STAT1 transcript (lane 2). Treating cells with ATRA with or without IFN-␥ caused large increases in similar fashion (lane 4). These results indicated that treating THP-1 cells with ATRA, with or without IFN-␥, increased significant gene expression of STAT1. It has been shown that STAT1 can be activated as a homodimer that moves into the nucleus and acts as a mature transcription factor by binding to the GAS element (39). The dimerization of STAT1 requires tyrosine phosphorylation of STAT1 in a manner triggered by IFN-␥ (38). Mutant STAT1 (STAT1-Y701Fm, replacing tyrosine 701 with phenylalanine) is unable to undergo the tyrosine phosphorylation dependent dimerization process (40). To test the possibility that IFN-␥ may be involved in activating STAT1 to up-regulate the ACAT-1 gene, we prepared wild-type STAT1 cDNA (STAT1-Y701) and the mutant STAT1 cDNA (STAT1-Y701Fm) in pRC/CMV vector. We then transfected these constructs in-dividually into the IFN-␥ and/or ATRA-treated THP-1 cells, and measured ACAT-1 P1 promoter activity. As shown in Fig.  9B, when cells were treated with IFN-␥ alone, a significant enhancement of the P1 promoter was seen when these cells were transfected with wild-type STAT1 cDNA. The enhancement was not seen when mutant STAT1 cDNA was used (comparing the sizes of the second bar in the STAT1-Y701 panel with the second bar in the STAT1-Y701Fm panel). These results imply that IFN-␥ is involved in stimulating the phosphorylation of STAT1, causing the dimeric form of STAT1 to bind to the GAS element. To further test this interpretation, we treated THP-1 cells with IFN-␥ alone, and transfected with or without the wild-type or the mutant STAT1 cDNAs, then prepared the nuclear extracts of these cells and performed EMSAs. As shown in Fig. 9C, the two GAS-specific bands were detectable in the nuclear extracts of cells transfected with wild-type  10 with lane 2). Results of the gel supershift assays using anti-STAT1 and anti-Sp1 antibodies confirmed that these two bands were complexes resulting from specific interactions of STAT1 and Sp1 with the ACAT-1 P1 promoter (lanes 8 and 9). Additional EMSAs, using the nuclear extracts from control (untransfected) THP-1 cells or from mutant STAT1 cDNA-transfected cells showed that the two bands described above migrated faster. These bands were not supershifted with anti-STAT1 antibodies, but were supershifted with anti-Sp1 antibodies (comparing lane 4 and 5 with 12 and 13). These results showed that the GAS site, not the Sp1 sites, formed specific complexes with the wild-type STAT1 after activation by IFN-␥ through the tyrosine-phosphorylation dependent mechanism. DISCUSSION ACAT-1 mRNAs and protein contents are significantly increased during the human monocyte-macrophage differentiation process in vitro (21,41). Its protein content is amply present in macrophage-derived foam cells localized in the human atherosclerotic lesion, implying that up-regulation of the ACAT-1 gene plays important roles in macrophage foam cell formation in atherosclerosis (42). In mouse macrophages, ACAT-1 message was found to be up-regulated by cells with IFN-␥ (12). The molecular basis of these findings has not been pursued at the gene transcription level. In our current work, we showed that IFN-␥ increased ACAT-1 mRNAs and protein contents during the human blood monocyte-macrophage differentiation process. We then found that treating the human monocyte-like THP-1 cells with ATRA and IFN-␥ caused upregulation of ACAT-1 gene expression in cell-type specific manner. To elucidate the molecular basis of this finding, we identified a 159-bp core region with Sp1 elements that is responsible for the P1 promoter activity. This region also contains an IFN-␥ activated sequence (GAS) that overlaps exactly with the second Sp1 element (TGGGCGGAA, with the Sp1 site underlined). To our knowledge, this is the first example in literature describing an overlapping Sp1/GAS site. Using luciferase constructs in transient trasfection studies, we demon- The wild-type and mutant GAS DNA fragments as indicated were labeled; 1 ϫ 10 4 dpm of labeled probe was used in each binding reaction. C, lane 1, 32 P-labeled wild-type fragment as the probe alone serving as negative control. Lane 2, binding reaction between labeled probe and the nuclear extracts. Lane 3, competition by adding 100-fold molar excess of cold wild-type probe; lanes 4 -6, supershift reactions by adding 1 l of anti-Sp1 antibody, 1 l of anti-STAT1 antibody, or 1 l of anti-Sp1 antibody and 1 l of anti-STAT1 antibody as indicated to the binding reaction; lanes 7-10, competition by adding 100-fold molar excess of probe containing mutation within the first, or second, or the first two, or all four Sp1 elements as indicated to the binding reaction; lanes 11 and 12, supershift reactions by adding 1 l of anti-Sp1 antibody or 1 l of anti-STAT1 antibody as indicated to the binding reaction described for lane 10. D, lanes 1-3, the same conditions as described for Fig. 7C, lane 1-3, were employed; lane 4, competition by adding 100-fold molar excess of probe containing the mutant GAS element; lanes 5 and 6, supershift reactions by adding 1 l of anti-Sp1 antibody or anti-STAT1 antibody as indicated to the binding reaction described for lane 4; lanes 7-9, the same conditions as described for lanes 1-3 were employed, except the mutant GAS DNA fragment was used as the labeled probe; lanes 10 and 11, supershift reactions by adding 1 l of anti-STAT1 antibody or anti-Sp1 antibody as indicated to the binding reactions described in lane 8. strated that the combination of IFN-␥ and ATRA is needed to enhance ACAT-1 P1 promoter activity. Additional experiments using RT-PCR and EMSA showed that ATRA caused large induction of the transcription factor STAT1, while IFN-␥ triggered the phosphorylation dependent activation of STAT1. The activated STAT1 then acts by binding to the overlapping GAS/ Sp1 site in the ACAT-1 P1 promoter. Our work dissects the non-sterol-mediated ACAT regulation at the transcriptional level, and provides a molecular mechanism to account for part of the effects of IFN-␥ in causing macrophage foam cell formation in vitro.
In atherosclerosis, the infiltration of T-cells and monocytederived macrophages into the intimal layer of the artery is believed to lead to foam cell formation. Activated T cells found in human atheroma secrete high levels of IFN-␥ (43,44). IFN-␥ has been shown to exert certain proatherosclerogenic actions in vitro. It induces VCAM-1 on endothelial cells (45), decreases apoE secretion, and increases uptake of hypertriblyderidemic very low density lipoprotein on macrophages (46), induces myosin heavy chain-II on macrophages and smooth muscle cells (47), and induces scavenger receptors on smooth muscle cells during atherogenesis (48). On the other hand, IFN-␥ has also been shown to exert protective action against atherosclerosis in certain in vitro systems examined (49,50). Recently, it has been shown that apoE knockout mice crossed with the IFN-␥ receptor knockout mice display reductions in lesion size, lipid accumulation, and cellularity (51). In addition, in mice, posttransplant graft arteriosclerosis is associated with the presence of IFN-␥; the serological neutralization or the genetic absence of IFN-␥ markedly reduces the extent of intimal expansion (52). These results support the notion that IFN-␥ is pro-atherogenic in vivo. If this concept holds true, then our finding described here may explain some of the effects of IFN-␥ on foam cell formation in vivo.
IFN-␥ exhibits antigrowth or antiproliferation effects in various target cells. Its effects often occur synergistically with retinoids (53). To cite a few examples, in various myelogenous leukemic cell lines, Gianni et al. (54) showed that ATRA can bypass IFN/IFN receptors and induce the expression of IFNregulated genes including STAT1; Matikainen et al. (55,56) showed that ATRA causes up-regulation of several IFN-specific transcription factors and signal inducers including STAT, and enhances their responsiveness toward IFNs. The molecular mechanism(s) for the synergism observed between ATRA and IFNs in these studies remain to be elucidated. Our work described here may serve to explain some of the synergistic actions of ATRA and IFNs described in these studies. FIG. 9. STAT1 is involved in the synergistic effect of IFN-␥ and ATRA. Total RNAs were prepared from THP-1 cells treated for 40 h with or without IFN-␥ (100 units/ml), or ATRA (10 Ϫ6 M), or IFN-␥ (100 units/ml) plus ATRA (10 Ϫ6 M) as indicated. A, quantitation by RT-PCR (30 cycles). Primers were used for STAT1 and the GAPDH cDNAs as described under "Experimental Procedures." Control experiments indicated that between cycles 25 and 35, the IFN-␥ receptor transcripts and the STAT1 transcript could be estimated semiquantitatively by RT-PCR (data not shown). B, THP-1 cells were co-transfected with the Luc construct containing the wild-type ACAT-1 P1 promoter core region, and the wild-type STAT1 (indicated as STAT1-Y701), or the mutant STAT1 (indicated as STAT1-Y701Fm), or the empty vector (indicated as control). 7 h after transfection, the cells were treated in various manners as indicated for 40 h. The luciferase activities of treated cell extracts were then determined in the same way as described in the legend to Fig. 1A. C, EMSAs using nuclear extracts from the transfected THP-1 cells treated for 40 h with IFN-␥ (100 units/ml). The wild-type 159-bp core region DNA was labeled as probe; 1 ϫ 10 4 dpm of labeled probe was used for each binding reaction. Lane 1, 32 P-labeled probe alone. Lane 2, binding reaction between labeled probe and the nuclear extracts. Lane 3, competition by adding 100-fold molar excess of cold probe. Lane 4, supershift reaction by adding 1 l of anti-Sp1 antibody. Lane 5, supershift reaction by adding 1 l of anti-STAT1 antibody. For lanes 2-5, nuclear extracts were from cells transfected with the empty vector. Lanes 6 -9 and 10 -13 are results using the same series of reaction conditions as described in lanes 2-5, but using the nucleic extracts from THP-1 cells transfected with wild-type STAT1 cDNA or with mutant STAT1 cDNA, respectively.