Adipocytic Differentiation and Liver X Receptor Pathways Regulate the Accumulation of Triacylglycerols in Human Vascular Smooth Muscle Cells*

(VSMC) is a feature of atherosclerotic plaques. In this study we describe two mechanisms whereby human VSMC foam cell formation is driven by de novo synthesis of fatty acids leading to triacylglycerol accumulation in intracellular vacuoles, a process distinct from serum lipoprotein uptake. VSMC cultured in adipogenic differentiation medium accumulated lipids and were induced to express the adipocyte marker genes adipsin, adipocyte fatty acid-binding protein, C/EBP (cid:1) , PPAR (cid:2) , and leptin. However, complete adipocyte differentiation was not observed as numerous genes present in mature adipocytes were not detected, and the phenotype was re-versible. The rate of lipid accumulation was not affected by PPAR (cid:2) agonists, but screening for the effects of other nuclear receptor agonists showed that activation of the liver X receptors (LXR) dramatically promoted lipid accumulation in VSMC. Both LXR (cid:1) and LXR (cid:3) were present in VSMC, and their activation with TO901317 resulted in induction of the lipogenic genes fatty acid synthetase,

Atherosclerosis is the principle cause of coronary artery disease in the Western world with plaque rupture leading to an acute coronary event (1). Fatty plaques are late stage atherosclerotic lesions consisting of a lipid and matrix core confined within the vessel wall by a fibrous cap. Stability of these lesions depends on the abundance and reparative capacity of the VSMC 1 in the fibrous cap (2). Lipid-filled VSMC are the dominant foam cell type in early lesions and are present at later stages of atherosclerosis (3,4); such a perturbation in VSMC biology could promote lesion development and contribute to plaque instability. Therefore, an analysis of the mechanisms of lipid accretion in VSMC may provide insights into the process of atherosclerosis and other conditions of vascular dysfunction.
Lipid filled "foam cells" within atherosclerotic lesions are derived from both VSMC and monocytes. Macrophage foam cells appear in juvenile fatty streaks and become most abundant in the fatty plaque. Infiltrating monocytes differentiate in the atherogenic environment of the vessel wall. This is accompanied by expression of receptors such as CD36 and scavenger receptor type A and lipoprotein uptake, resulting in the formation of foam cells rich in cholesterol and cholesterol esters (CE) (5,6). Lipoprotein uptake by VSMC has been reported (7)(8)(9). However, unlike the macrophage situation, establishment of a scavenger receptor phenotype is not essential for VSMC lipid uptake. VSMC do not express high levels of scavenger receptors in atherosclerotic tissue and are resistant to lipid accumulation after exposure to normal LDL and oxidized LDL in vitro (10 -14). The absence of macrophage scavenger receptors in VSMC indicates that other mechanisms of lipid accumulation may occur. These could include the utilization of LDL and VLDL receptors in the uptake of aggregated LDL and ␤-VLDL, a process that occurs in vitro (15,16), or the uptake of nonesterified fatty acids from serum (17,18). This mechanism of VSMC foam cell formation is of particular interest due to the elevated levels of FA in diabetes and the association of this condition with atherosclerosis (17). An alternative mechanism would be de novo synthesis of lipid, and it has long been known that VSMC, human arterial tissue, and foam cells incubated with radioactive acetate synthesize phospholipids, triacylglycerols (TG), and CE (19,20), the type of lipid produced being dependent on VSMC phenotype and effected by oxysterols (21,22). Excessive de novo synthesis could result in lipid accumulation and foam cell formation. Furthermore, the lipid in dis-eased tissue is thought to be derived from both cellular and serum lipoprotein sources. CE and TG that accumulate in fatty streaks and advanced lesions are rich in oleic acid; this may in part result from cellular synthesis (23)(24)(25).
De novo lipogenesis and TG storage are specialized functions of white adipose tissue. The capacity of adipocytes to regulate lipid metabolism is associated with terminal differentiation of preadipocytes, a process generally linked with expression of the transcription factors PPAR␥, C/EBP␣, SREBP1/ADD1 (26 -28). VSMC are known to have mesenchymal characteristics; they are not terminally differentiated and show considerable phenotypic plasticity in normal and diseased tissues (29 -31). We, therefore, hypothesized that VSMC maintain the propensity to develop an adipocyte phenotype if exposed to adipogenic conditions and that this would result in lipid accumulation. Enhanced de novo lipogenesis is also observed in non-adipocytic cells, in pancreatic ␤-cell lipotoxicity in type 1 diabetes, renal cell lipotoxicity, and alcohol induced fatty liver (32)(33)(34)(35). In these situations SREBP1 plays a central role in promoting de novo lipogenesis (32)(33)(34)(35). By using a combination of lipid analysis, gene expression data, and nuclear receptor agonists, we demonstrate that VSMC can synthesize and store considerable amounts of TG. The generation of lipid-filled VSMC resulted from either adipocytic differentiation or direct promotion of lipogenesis as the result of LXR/SREBP1c activation. In addition we identify VSMC expressing high levels of FASE, SREBP1, LXR␣, and adipsin within human atherosclerotic plaques, suggesting that de novo lipogenesis may play a role in VSMC lipid accumulation in vivo.
Lipids were extracted from these homogenized arterial specimens and from cultured VSMC using chloroform-methanol by our modification of the method of Bligh and Dyer with the inclusion (for VSMC) of a brief initial sonication on ice (37). Lipids were separated on Whatman PK6F silica gel thin layer chromatography plates (20 ϫ 20 cm; 1-mm thickness) eluted in a mixture of hexane, diethyl ether, and glacial acetic acid (70:30:1 by volume) with authentic lipid standards run alongside, and stained with iodine vapor. Fatty acids (FA) were liberated from TLC-purified lipid fractions by saponification followed by acidification, extraction into diethyl ether, and derivatization to methyl esters (38). The individual FA species in each fraction were identified and quantified by GC using a Vega 6130 gas chromatograph (Carlo Erba Instruments, Milan, Italy) with a DB-1 fused silica capillary column (J&W Scientific, Folsom, CA) and flame ionization detector as described previously (38). Peak areas were measured using a Chrom-Card data system (Carlo Erba Instruments).
For Oil Red O (ORO) staining, VSMC were fixed with 4% formaldehyde in phosphate-buffered saline and then stained for 20 min with 0.05% ORO in isopropanol and water (3:2 by volume). Cells were then washed with water and viewed by microscopy.
RNA Extraction and RT-PCR-RNA was isolated from VSMC using RNA STAT-60 as described by manufacturer (Tel-Test Inc, Friendswood, TX). Human preadipocyte and adipocyte RNA was provided by Dr. C. Sewter and derived from fatty tissue stromovascular cells cultured in growth medium or serum-free Dulbecco's modified Eagle's medium/Ham's F-12 adipose differentiation medium (39,40). For the reverse transcription total RNA from human VSMC cultures was subjected to "Super-RT" avian myeloblastosis virus reverse transcriptase (HT Biotechnology Ltd, Cambridge, UK). Experimental conditions for the RT-PCR were described previously (31) except that random hexamers (Promega, Madison, WI) were added to the RT step as well as oligo-dT. ␤-Microglobulin was used to control for RT-PCR (n ϭ 3). Most of the PCR primer pairs were designed to be functionally active at 58°C annealing temperature and to have a product size between 200 and 600 bp. Preliminary experiments with a range of cycle numbers were performed to ensure that PCR amplification was within the "log phase" of the reaction. The functional activity of the primers and size of positive control amplicons were determined using RNA from pre-adipocytes, adipocytes, and if necessary atherosclerotic tissue (data not shown). Primer sequences are available on request. The PCR products were subjected to electrophoresis on a 1% agarose gel in the presence of ethidium bromide and quantified by electronic digital imaging using a MultiImage light unit with Alphamager 1200 software (Alpha Innotech Corp, San Leandro, CA). The SREBP1 and PPAR␥ primers do not distinguish between SREBP1a/c or PPAR␥1/␥2.
Western Blotting-Cell extracts were prepared by lysis in non-reducing Laemmli sample buffer minus bromphenol blue, and protein was quantified by the BCA assay (Pierce). Dye and reducing agent was added to the samples, and the proteins were separated by SDS-PAGE and transferred to Immobilon-P membrane (Millipore, Billerica, MA). Equal protein loading and transfer were confirmed by Coomassie and Ponceau S staining. The membrane was blocked with 10% milk, phosphate-buffered saline, 0.1% Tween 20 for 1 h and then incubated with diluted primary antibody in 3% milk, phosphate-buffered saline, and 0.1% Tween 20. Antibodies reactive toward LXR␣ (⌯114, sc13068), LXR␤/UR (N20, sc1001), SREBP1 (H160, sc8984) were from Santa Cruz (Santa Cruz Biotechnology, Santa Cruz, CA). Antibody reactive to human fatty acid synthetase (905-069) was from Assay Designs (Assay Design Inc, Ann Arbor, MI). Specificity of reactivity for the LXR␤ antibody was confirmed by competition with excess peptide (sc1001p, OR1 N20) as described by the manufacturer (Santa Cruz). The signal was detected using a 1:5000 dilution of horseradish peroxidase-conjugated secondary antibody and revealed with enhanced chemiluminescence (Amersham Biosciences).
Reporter Plasmids and Mutation of LXREs in the SREBP1 Promoter-pFASprom2000-Luc was obtained from Professor J. Tavare. 2 kb of the murine FASE promoter was cloned into pGL3-basic (41). SREBP1c reporter plasmids pBP1c2600-Luc and pBP1c357-Luc contained 2.6 kb and 357 bp of the murine SREBP1c promoter cloned into pGL2-basic, kindly provided by Dr. H. Shimano (42). Both SREBP1c promoter plasmids have the oxysterol-inducible region that contains two LXR binding sites (LXREa and -b). LXREa and LXREb were mutated to non-binding sites using the following oligonucleotides: mutant LXREa, Ϫ265 to Ϫ306, tgcgccagcgcgcgctggggAtTctggcggAcTctgtcgtcc; mutant LXREb, Ϫ264 to Ϫ223, gacgcggttaaaggcggaAgTccgctagAaTccccggcccca. The oligonucleotides were annealed to the wild type plasmids, and the DNA was filled in by two-temperature step PCR using Pfu1 polymerase. After extension the ends of the PCR products were treated with kinase, ligated, and transfected into Escherichia coli. The resulting plasmids were sequenced to ensure mutation of both LXREa/b and a correct DNA sequence at the 5Ј junction of the two oligonucleotide binding sites. The mutated SREBP1c promoter fragments were cut out and cloned into the parental pGL2-basic plasmid to remove the possibility of mutations in the luciferase gene. pRL-SV40 control plasmid (Promega) encodes Renilla luciferase regulated by the SV40 promoter.
Transient Transfection and Luciferase Reporter Assays-VSMC were plated at a density of 1-5 ϫ 10 5 per dish, and cultured for 24 h before plasmid transfection. A total of 5 g of luciferase reporter plasmid and 0.1 g of the pRL-SV40 Renilla luciferase control (Promega) was mixed with 25 l of Superfect for each dish as described by the manufacturer (Stratagene, La Jolla, CA). Cells were transfected for 2 h, washed to remove plasmid, and left overnight in growth medium to recover. New culture medium with agonists was added for 48 -72 h. Six dishes were analyzed for each culture condition. Cells were scraped off, spun down, and lysed in 50 l of cell culture lysis reagent (Promega). 20 l of lysate was analyzed using the dual luciferase reporter assay system according to manufacturer's instructions (Promega), and luciferase activity was detected using a luminometer (EC&G Berthold, Bad Wildbad, Germany). Transfection efficiency variations within each assay run were normalized for Renilla luciferase emission. The data were reported as relative luciferase units, representing the mean of the recorded firefly luciferase emission values minus the low background from non-transfected VSMC.

Accumulation of ORO Staining in VSMC Cultured in Adipo-
cyte Differentiation Medium-ORO staining was low or absent from VSMC cultured in growth medium. In contrast some VSMC isolates cultured in adipocyte DM accumulated considerable amounts of ORO lipid. Analysis of a responsive VSMC isolate show that the abundance of cells containing lipid and the amount of lipid per cell was enhanced with increased periods of culture in adipogenic conditions; this was not observed when the cells were grown in 20% FCS (Fig. 1, A and B). Small clear vacuoles were observed in VSMC cultured in growth medium and DM. Lightly ORO-stained, small highly refractive intracellular droplets would then develop, often associated with stress fibers. The abundance, size, and staining of the vacuoles increased in cells cultured with DM, with the result that some cells were heavily loaded with lipid after 3 weeks of culture. The large globular lipid vacuoles were morphologically indistinguishable from those observed in true adipocyte cultures. The extent of lipid accumulation was dependent on the VSMC isolate, some of which did not accumulate lipid when cultured in DM. The variability of VSMC lipid accumulation is demonstrated by the photographs in Fig. 1C; the three VSMC isolates were cultured for 21 days in DM and stained with ORO.
TLC analysis revealed that TG was the major lipid type accumulated in VSMC cultured in DM. Fig. 1D shows the fatty acid composition of the TG from cells cultured in growth medium or DM, with the proportion of oleic acid (C18:1) greater in the TG-derived from VSMC cultured in DM. VSMC cultured in DM accumulated TG enriched with oleate, the same lipid type that is synthesized and stored in white adipose tissue and 3T3 cells induced to undergo adipocyte differentiation (43)(44)(45).
Adipocyte Marker Genes Are Up-regulated in VSMCs Cultured in DM-Semi quantitative RT-PCR was used to investigate expression in VSMC of a number of gene markers known to be up-regulated during adipocyte differentiation. Fig. 2A shows that expression of adipsin, adipocyte fatty acid-binding protein, C/EBP␣, and PPAR␥ were elevated in VSMC cultured in DM; gene expression was low or not detectable from cells cultured in growth medium. The phenotype was reversible, as shown by the reduction in marker gene expression after replacement of DM with medium containing 20% FCS (DM/FCS). ORO staining of the cells also decreased as a result of changing culture conditions from DM (21 days) to growth medium (data not shown). Expression of FASE, the key gene involved in FA synthesis, SREBP1, and leptin was also enhanced as the result of VSMC culture in DM (Fig. 2B). The early adipocyte marker adipose differentiation-regulated protein (adipophilin) and the leptin receptor were expressed at equivalent levels in VSMC cultured under normal or DM conditions; they both reportedly increase during true adipocyte differentiation (46,47), Finally, certain genes known to be up-regulated in mature adipocytes were not detected in VSMC under any of the culture conditions tested. Hormone-sensitive lipase, glycerol-3-phosphate dehydrogenase (G3PDH), aPM1 (adipose most abundant gene transcript 1 (adiponectin)), and the lipid vacuole surface protein perilipin show abundant expression in the mature adipocyte but were not detected in VSMC cultured in DM (Fig. 2C).
Effect of Nuclear Receptor Agonists on Lipid Accumulation in VSMCs-The above experiments indicated that VSMC could adopt an adipocyte phenotype in response to DM; therefore, we investigated whether specific agonists for members of the PPAR and LXR families of nuclear hormone receptors could enhance this differentiation. VSMC were treated with 5 or 10 M PPAR␥ agonist (BRL49653/rosiglitazone) in DM, serumfree conditions, or growth medium. No change in the rate of lipid accumulation was observed (as measured by ORO staining) between cells treated with the vehicle Me 2 SO or BRL49653. Treatment of VSMC with ciglitazone also had no effect on promoting lipid accumulation (not shown). Agonists for PPAR␣ (WY14643) and PPAR␤/␦ (GW610742) were also tested under the above culture conditions. WY14643 did not increase the rate of lipid accumulation in VSMC, whereas treatment with the GW610742 slightly enhanced the rate of ORO accumulation in cells cultured in DM (not shown).
In contrast to the effect of the PPAR agonists, treatment of the VSMC isolates with the LXR agonist TO901317 dramatically increased the rate of lipid accumulation. Cells treated with 0.5-10 M TO901317 in serum-free medium, DM, or me-dium containing 2.5% FCS, 5% FCS, 10% lipoprotein-deficient charcoal stripped bovine serum stained positively for ORO. TO901317 also promoted lipid accumulation in the VSMC isolate 99.2.5.12A (Fig. 3A); this isolate had otherwise been shown to be resistant to lipid accumulation even when cultured in DM. Lipid accumulation was clearly detectable after 5 days of culture with TO901317. TLC showed that TG was increased in VSMC treated with TO901317, with elevations in content of palmitic acid (C16:0) and, more markedly, of oleic acid (C18:1) (Fig. 3B).
LXR␣ and LXR␤ Are Expressed by VSMC and Can Activate Promoters via the LXRE-The LXR nuclear receptor subfamily consists of two members, LXR␣ (NR1H3) and LXR␤ (NR1H2) (48). Western blot analysis of cell lysates from four VSMC isolates with an antibody specific for LXR␣ detected a 55-kDa protein, consistent with the size of LXR␣ in other cells types (Fig. 4A). Using an LXR␤ antibody, two bands, a minor band at 75-kDa and a 105-kDa protein, were detected (Fig. 4B). Specific antibody binding for both bands was confirmed by competition with peptide (not shown).
The SREBP1c promoter is known to be activated by LXR (42, 49 -51). Western blot analysis of VSMCs cultured with or without LXR agonist indicated that SREBP1 protein levels were elevated after TO901317 treatment, the increase being greatest when the cells were cultured in low serum (2.5% FCS) or in DM (Fig. 4C). A slight increase in SREBP1 expression was observed in TO901317 treated cells cultured in growth medium (20% FCS).
To confirm that LXR activated genes in VSMCs via the LXRE, SREBP1c promoter-luciferase reporter constructs with intact or mutated LXR binding sites (LXREa/b) were tested for responsiveness to TO901317. The 350-bp and 2.6-kb murine SREBP1c promoters were activated by TO901317 but not by vehicle Me 2 SO (Fig. 4D). Mutation of both LXRE sites eliminated TO901317-dependent activation in the 350-bp promoter and dramatically reduced activation of the larger 2.6-kb promoter.
TO901317 Treatment Results in the Activation of the FASE, SREBP1c, and ABC-A1 in VSMCs-The level of expression and activity of FASE regulate the rate of FA synthesis. Western blotting demonstrated that FASE was dramatically increased FIG. 2. Expression of adipocytic marker genes in human VSMC. A, RNA derived from VSMC cultured in growth or differentiation conditions was used for quantitative RT-PCR using primers specific for adipsin, adipocyte fatty acid-binding protein (AFBP), C/EBP␣, and PPAR␥. Expression levels of the ␤-microglobulin demonstrate that the amount of VSMC RNA was the same between the three culture conditions. Cells were cultured for 21 days in growth medium containing 20% FCS (FCS), for 21 days in DM followed by 3 days culture in growth medium (DM⁄FCS), or for 21 days in adipocyte differentiation medium (DM). RNA from human pre-adipocytes and differentiated adipocytes act as controls for primer function and PCR product size (fourth and fifth lanes). The amount of adipose RNA was less than for the VSMC to prevent saturation of the RT-PCR reaction. B, RT-PCR analysis of FASE, SREBP1, leptin, adipose differentiation-regulated protein (adipophilin) (ADRP), and leptin receptor (OB-R) gene expression in VSMC cultured in conditions mentioned above. All these genes have been associated with adipocyte function but are not regarded as specific markers for the adipocyte phenotype. The SREBP1 primers recognize both the 1a and 1c isoforms. The absence of certain adipocyte markers in cultured VSMC is shown in C. Expression of hormone-sensitive lipase (HSL), glycerol-3-phosphate dehydrogenase (G3PDH), aPM1 (adipose most abundant gene transcript 1 (adiponectin)), and perilipin genes were detected by RT-PCR in the control mature adipocyte RNA; they were not readily detectable in RNA derived from the VSMC cultured in normal or differentiation conditions. Extremely low levels of hormone-sensitive lipase and perilipin were noted in a few reactions for differentiated VSMC.
in VSMC after TO901317 treatment, as was expression of the 150kDa SREBP1c band. Levels of the 160-kDa protein, probably the SREBP1a isoform, were not enhanced by TO901317 (Fig. 5B). This would be consistent with reports that LXR binds and activates the SREBP1c promoter but not the SREBP1a promoter (42,49) and that FASE gene transcription is activated by both LXR and SREBP1 (52). Low molecular weight cleavage products of SREBP1 were also enhanced by TO901317. Protein levels of LXR␤ were not influenced by treatment of VSMC with TO901317 or the other nuclear receptor agonists (Fig. 5C).
Activation of the FASE and SREBP1c promoters by TO901317 are shown in Fig. 5E. VSMC transfected with the FASE or SREBP1c promoter luciferase reporters were cultured with vehicle Me 2 SO or 5 M TO901317. There was significant activation of both the FASE promoter and SREBP1c after TO901317 treatment. (53)(54)(55)(56)(57). We were interested in whether 27-hydroxycholesterol (27-HC), an oxysterol synthesized by normal VSMC and also present in diseased vessels (58), was an LXR agonist. 27-HC did not significantly activate the SREBP1c luciferase reporter plasmid pBP1c357 in VSMC (Fig. 6A), and this was confirmed in a second VSMC isolate (data not shown). In contrast, the oxysterol 24(S)25-EC was an efficacious LXR agonist in VSMC (Fig. 6A). Western blot analysis for SREBP1 confirmed that 27-HC was not an LXR agonist in VSMC (Fig. 6B), in contrast to the enhanced expression of SREBP1 protein with 24(S)25-EC, TO901317, or the RXR ligand, 9-cis-retinoic acid.

Influence of Oxysterols 27-Hydroxycholesterol and 24(S)25-Epoxycholesterol on LXR Function in VSMC-Specific natural oxysterols have been shown to be physiological ligands for LXR
27-HC Is an Antagonist for TO901317 and 24(S)25-Epoxycholesterol-mediated LXR Activation in VSMC-Next we inves- . The cells were cultured in growth medium containing 20% FCS, medium with 2.5% FCS, or adipocyte differentiation medium. D is the luciferase activity of the wild type (wt) or LXREmutated SREBP1c promoters in human VSMC; the transfected cells were treated with vehicle Me 2 SO (Ϫ) or 10 M TO901317 (ϩ). The luciferase reporter plasmid pBP1c357-Luc (SREBP1c-360luc) and pBP1c2600-Luc (SREBP1c-2.6luc) contained 357-bp and 2.6-kb fragments of the SREBP1c promoter. VSMC transfected with plasmid were harvested 72 h after treatment, and the cell lysate was assayed for luciferase activity. RLU, relative luciferase units. tigated whether 27-HC could influence the ability of TO901317 or 24(S)25-EC to activate the SREBP1c luciferase reporter plasmid pBP1c357-Luc. The addition of 27-HC resulted in a reduction in TO901317-induced SREBP1c promoter activity (Fig. 6C). 27-HC also repressed promoter activation by 24(S)25-EC. There was a slight increase in promoter activity when VSMC were treated with 1, 5, or 10 M 27-HC in the presence of the Me 2 SO vehicle (Fig. 6C).
Oleic Acid Was Enriched within the Triacylglycerol Purified from Atherosclerotic Lesions-GC was used to determine the fatty acid composition of TG from advanced lesions (plaques) and adjacent macroscopically normal arterial tissue from three carotid arteries. Palmitate (C16:0) and oleate (C18:1) were the dominant FA species within the TG from both normal and plaque tissue. There was significant enrichment for oleate within plaque TG (Fig. 7). The fatty acid composition of the TG isolated from normal vessel (Fig. 7) was similar to that of VSMC cultured in growth medium (Fig. 1D), whereas the fatty acid composition of TG from advanced lesions (Fig. 7) resembled those of VSMC cultured in DM or with TO901317 (Figs. 1D and 3B). The CE content of the arterial samples was dramatically higher than for the cultured VSMC. In the vessel CE was far more abundant than TG and characterized by a high proportion of the essential FA linoleate (C18:2).
VSMCs within Atherosclerotic Plaques Express High Levels of Adipsin, FASE, SREBP1, and LXRa-Immunohistochemis-try was used to determine expression levels of phenotypic markers and transcription factors in normal and diseased human arteries (Fig. 8). Medial VSMCs in normal aortic vessel showed no reactivity to antibodies recognizing adipsin and the lipogenesis enzyme FASE (a and b). Similarly, the transcription factor SREBP1 was not detected, whereas LXR␣ was observed at low levels (c and d). The expression of these markers and transcription factors was markedly up-regulated in VSMC from diseased vessel. Medial VSMCs located beneath regions of atherosclerotic plaque were positive for adipsin (e), FASE (f), SREBP1 (g), and LXR␣ (h). The medial cells were located in a region of lesion that weakly stained with ORO (n). Although some of these cells were lipid-filled (8eЈ), others were not (f-h). In addition, intimal VSMCs deep within the lipid-rich OROpositive regions of the plaque stained positively for adipsin, FASE, SREBP1, and LXR␣ (i-l). Many of these cells were morphologically foam cells with obvious cytoplasmic lipid vesicles. Foam cells of VSMC and macrophage origin were present in the intimal region, and both stain for SREBP1 (k). o shows extensive ORO staining in the lipid-rich intimal region of the plaque. Normal vessel was negative for ORO staining (not shown). DISCUSSION Lipid-containing VSMC within atherosclerotic plaques express elevated levels of the adipocyte marker adipsin and genes FIG. 5. Treatment of human VSMC with TO901317 up-regulates expression of both LXR-and SREBP1c-responsive genes. A, an antibody recognizing FASE was used to Western blot whole cell lysates from VSMC cultured for 6 days in the presence of vehicle (Me 2 SO (DMSO)), the nuclear receptor agonists for PPAR␥ (BRL49653), LXR␣/␤ (TO901317), PPAR␦ (GW610742), or TO901317/BRL49653 and GW610742/ BRL49653 combinations. Each agonist was added at 0.5 and 5 M to the VSMC cultures. Western blots shown in B and C used antibodies reactive toward SREBP1 and LXR␤, respectively. The protein lysate loaded in each lane was the same as mentioned above for A. In blot B the dominant protein band of size 150 kDa was SREBP1c; the weak band at 160 kDa was probably SREBP1a. The low size bands are proteolytically cleaved SREBP1. D, RNA derived from VSMC cultured with or without nuclear agonist was used for quantitative RT-PCR using primers specific for SCD-1, ABC-A1, SREBP1a/c, FASE, and ␤-microglobulin. Cells were cultured in medium plus 5% lipoprotein-deficient charcoal-stripped bovine serum, with the exception of lane 5, where growth medium was used (20% FCS). Panel E shows results from luciferase reporter assays demonstrating the level of expression of the FASE or SREBP1c promoter in VSMC. VSMC transfected with pFASprom-2000 (FASE 2.0-kb Luc) or pBP1c2600-Luc (SREBP1c 2.6-kb Luc) were treated for 72 h with either the vehicle Me 2 SO (Ϫ) or 5 M TO901317 (ϩ). Significant enhancement of both FASE and SREBP1c promoter activity was observed with the LXR agonist TO901317 (n ϭ 6).
involved with lipogenesis, including FASE, SREBP1, and LXR␣. In vitro, elevated de novo fatty acid synthesis in VSMC leads to TG accumulation, suggesting that synthesized lipid may contribute to foam cell formation in vivo. In support of this, TG levels increase in vessels with age and atherosclerosis, albeit to a lesser degree than CE. Also noteworthy is that CE derived from intimal regions with abundant foam cells are oleate-rich, further supporting the view that lipid is derived at least partly from a cellular origin rather than exclusively from plasma lipoproteins (23)(24)(25). The preferential synthesis in vitro of TG rather than CE may reflect a characteristic function of VSMC, as suggested by TG accumulation after VSMC uptake of non-esterified FA (17,18). Alternatively, the serum-free or low serum culture conditions may restrict ACAT1 activity by limiting cholesterol substrate.
Adipocyte Differentiation of VSMC-In vitro certain VSMC isolates were predisposed to accumulate TG and develop an adipocyte phenotype when incubated in DM, with some expressing the late differentiation marker CEBP␣ (26,28,59). However, VSMC were generally unable to express the full range of adipocyte markers, and the phenotype was reversed when cells were transferred back to growth medium.
The conversion of mesenchymal stem cells, mice embryonic fibroblasts, and bone marrow stromal cells among mesenchymal phenotypes, in particular the osteoblastic and adipocytic lineages, is well established and regulated in part by expression of the lineage-specific transcription factors Cbfa-1 (core binding factor ␤ (Runx2)) and PPAR␥ (60 -62). VSMC also exhibit phenotypic plasticity, being convertible to an osteochondrocytic phenotype in vivo and in vitro (29 -31). Rosiglitazone did not promote adipocyte differentiation in VSMC. This suggests that PPAR␥ is not the central regulator for adipocytic conversion of VSMCs, similar to omental preadipocytes and certain murine stromal cell lines that are also refractory to PPAR␥ agonists (39,40,(61)(62)(63). Events promoting VSMC adipogenesis in vivo are unknown but presumably involve responses to factors in the diseased vessel wall. These could include deregulated transforming growth factor ␤/bone morphogenic proteins signaling, atherogenic diet, modified LDL, FA, and aging, all of which can promote lipid accumulation or adipocytic characteristics in bone marrow stromal cells (60,64,65).
Liver X Receptor and SREBP1 Pathway-Direct ligand activation of LXRs in VSMC up-regulated FASE, SREBP1c, and SCD-1, thus promoting TG accumulation via de novo FA synthesis and SCD-1-mediated ⌬ 9 desaturation. The increase in mature and processed SREBP1c together with SCD-1 expression indicated direct SREBP1c involvement in promoting TG synthesis (51,66). The identification of LXR␣ in VSMC was of particular interest because it was previously thought to be restricted to cells with high lipid metabolism (48,67). LXR␣ self-regulates its expression in sterol-loaded macrophages and is more effective than the ubiquitously expressed LXR␤ at Lipids isolated from diseased (advanced atherosclerotic lesions) and adjacent "normal" arterial tissue from carotid arteries of three patients (2 male, 1 female) were analyzed by TLC and gas chromatography. Palmitic (C16:0) and oleic acid (C18:1) were prevalent in the TG fraction from all samples, with significant enrichment for oleic acid in lesions. The fatty acid composition of CE from lesions is shown for comparison and was characterized by an abundance of linoleic acid (C18:2). promoting lipogenesis in adipocytes and hepatocytes (68 -73).
LXR activation by TO901317 promoted expression of genes necessary for lipogenesis, but not C/EBP␣, indicating that LXR activation alone was insufficient to promote the transcriptional cascade responsible for adipocyte differentiation. Indeed, there are conflicting reports of the ability of LXR agonists to promote adipogenesis in 3T3 cells and preadipocytes, although treatment resulted in lipid accumulation (74 -76). Overexpression of SREBP1c also inhibits adipogenesis in mice (77,78). Although the etiology of adipocyte differentiation and LXR activation pathways are distinct, they intersect at the point of SREBP1mediated activation of lipogenesis genes. Thus, within atherosclerotic lesions, factors that influence SREBP1 gene expression and cleavage might dramatically influence rates of fatty acid synthesis. Homocysteinemia and alcoholism, known risk factors for atherosclerosis, reportedly up-regulate the expression of SREBP-responsive genes in liver and VSMC (33,79). Hyperinsulinemia also elevates SREBP-1c activity, resulting in TG deposition in non-adipose tissues and lipotoxicity and, by analogy, may also play a role in vascular dysfunction and enhanced susceptibility to atherosclerosis in diabetes (34).
Oxysterol Involvement in Foam Cell Formation-Certain oxysterols are natural LXR ligands. 27-HC, an abundant oxysterol in atherosclerotic lesions, acted as an antagonist of LXR activity in VSMC in the presence of strong agonists. Reports of the LXR-activating ability of 27-HC in other cell types are conflicting, whereas in a cell-free system 27-HC acted as a partial agonist (53,56,57). The inability of 27-HC to act as an agonist in the present study may stem from factors such as the transcriptional co-regulator composition of VSMC or that very weak LXR agonists are not detected in our system. 27-HC was reportedly excluded from liver nuclei, unlike 24(S)25-EC and 24-HC (80). Should this be the situation in VSMC, then 27-HC would not be available to activate nuclear LXR, and the antagonist action observed would be independent of LXR binding.
In human advanced atherosclerotic lesions, mean levels of 27-HC are ϳ0.3 g/mg of wet weight or 0.7 mM, well above the concentrations that antagonized LXR in vitro (58,(81)(82)(83). Macrophages within advanced lesions express high levels of sterol 27-hydroxylase and are the primary source of 27-HC and derivatives (84 -86). The influence of 27-HC on lesion VSMCs would, therefore, be dependent on transfer of free 27-HC from macrophages to VSMC. The availability of 27-HC may also be modified by esterification and by the presence of cholesterol or other lipids in the lesion. 27-HC can be further metabolized by sterol 27-hydroxylase to form water-soluble cholestenoic acid (3␤-hydroxy-5-cholestenoic acid). Cholestenoic acid has been reported as an LXR agonist (55) or, conversely, an insignificant (56). In lesions cholestenoic acid levels are less than 3% that of those of 27-HC (87). The physiological relevance of cholestenoic acid as an LXR agonist needs to be determined.
Interestingly, sterol 27-hydroxylase expression by VSMC in the normal artery is greater than in VSMC in advanced lesions (84). 27-HC synthesis by VSMC in the normal vessel wall may exert an anti-atherogenic effect by restricting superfluous fatty acid synthesis as well as aiding cholesterol clearance. In humans, sterol 27-hydroxylase deficiency (a condition termed cerebrotendinous xanthomatosis) is invariably associated with premature atherosclerosis; this is thought to be due to diminished removal of cholesterol from macrophages (85). The role of VSMC in this accelerated atherosclerosis has not been studied.
Lipogenesis in VSMCs-This study provides evidence that lipid accumulation in VSMC in vivo can involve de novo lipogenesis, a process distinct from the cellular uptake of lipoproteins and FA from plasma. Identifying the factors in the plaque that induces the adipocytic/lipogenic phenotype in VSMCs and determining whether this phenotypic change occurs early in atherogenesis are important areas for further investigation. LXR agonists, as yet unidentified, may be present in atherosclerotic lesions, whereas factors that influence SREBP1 activity may include oxidative stress, cell senescence, and aging. Finally, the detrimental effects of enhanced VSMC lipogenesis and lipid accumulation should be considered when developing nuclear receptor agonists as therapeutic agents for atherosclerosis (88,89) FIG. 8. VSMC in atherosclerotic plaques express proteins associated with adipocyte differentiation and enhanced de novo lipogenesis. Medial VSMCs from the aorta (Fig. 8, a-d, m) do not express adipsin (a), FASE (b), or SREBP1 (c), as indicated by lack of brown staining. VSMC with nuclei counterstained with hemotoxylineosin are indicated by arrows in a-c. The negative control (m) lacked primary antibody. LXR␣ antibody weakly stained normal medial VSMCs brown (d); the sections were co-stained for ␣-SM actin shown as blue cytoplasmic staining. Diseased medial VSMCs beneath atherosclerotic plaques (Fig. 8, e-h, n) expressed high levels of phenotypic marker and transcription factor proteins. Adipsin, FASE, SREBP1, and LXR␣ antibody staining is brown, whereas ␣-SM actin staining is blue (e-h); representative positive VSMC are indicated by arrows in f-h). These medial cells from diseased carotid vessels are located in a region of weak ORO staining as shown at low power (n). Some ORO staining VSMC did exist at this location (eЈ). Analysis of VSMC in the lipid-rich region of the plaque (Fig. 8, i-l, o) also revealed strong antibody staining (i-l). The arrows show intimal VSMC foam cells expressing both the VSMC specific protein ␣-SM actin (shown as blue) and the indicated markers (shown as brown stain). Macrophage foam cells also expressed some markers (but not ␣-SM actin), as indicated by the arrowhead in k. ORO staining of the lipid-rich region is depicted in o.