The Peroxisome Proliferator-activated Receptor Promotes Lipid Accumulation in Human Macrophages*

The peroxisome proliferator-activated receptors (PPARs) are a family of fatty acid-activated transcription factors which control lipid homeostasis and cellular differentiation. PPAR (NR1C1) controls lipid oxidation and clearance in hepatocytes and PPAR (NR1C3) promotes preadipocyte differentiation and lipogenesis. Drugs that activate PPAR are effective in lowering plasma levels of lipids and have been used in the management of hyperlipidemia. PPAR agonists increase insulin sensitivity and are used in the management of type 2 diabetes. In contrast, there are no marketed drugs that selectively target PPAR (NR1C2) and the physiological roles of PPAR are unclear. In this report we demonstrate that the expression of PPAR is increased during the differentiation of human macrophages in vitro. In addition, a highly selective agonist of PPAR (compound F) promotes lipid accumulation in primary human macrophages and in macrophages derived from the human monocytic cell line, THP-1. Compound F increases the expression of genes involved in lipid uptake and storage such as the class A and B scavenger receptors (SRA, CD36) and adipophilin. PPAR activation also represses key genes involved in lipid metabolism and efflux, i.e. cholesterol 27-hydroxylase and apolipoprotein E. We have generated THP-1 sublines that overexpress PPAR and have confirmed that PPAR is a powerful promoter of macrophage lipid accumulation. These data suggest that PPAR may play a role in the pathology of diseases associated with lipidfilled macrophages, such as atherosclerosis, arthritis, and neurodegeneration.

Peroxisome proliferator-activated receptors (␣, ␦, and ␥) are ligand-activated transcription factors that control lipid and glucose homeostasis (1). They are members of the nuclear receptor family and form obligate heterodimers with the retinoid X receptor (2). PPAR␣ 1 is highly expressed in the rodent liver, where its activation up-regulates ␤-oxidation and thus promotes lipid clearance. PPAR␣ agonists, such as the fibrates, are effective lipid-lowering drugs. PPAR␥, which is activated by 15 deoxy-⌬ 12,14 PGJ 2 and the thiazolidinedione class of insulinsensitizing drugs, is expressed particularly in adipose tissue, where it initiates the differentiation cascade (3,4). Among its known target genes are adipocyte fatty acid-binding protein and fatty acid synthase, which are effectors of lipid accumulation during adipogenesis.
PPAR␥ has also been studied for its role in the formation of the atherosclerotic plaque, a lesion consisting of an accumulation of lipid-laden macrophages within the intima of the arterial wall. These studies started with the observation that PPAR␥ up-regulates CD36 (scavenger receptor-class B) expression in macrophages. This facilitates uptake of modified plasma low density lipoproteins into the macrophages potentially leading to the formation of foam cells (5,6). However, plaque formation is also appreciated to be an inflammatory response and it has been shown that activators of PPAR␥ inhibit the production of inflammatory cytokines such as tumor necrosis factor-␣, interleukin-6 and interleukin-1␤ (7)(8)(9)(10)(11). Although the PPAR dependence of these effects is controversial (12,13), the mechanism of this inhibition appears to be via repression of AP-1, NF-B, and STAT-1 activity (8,14,15).
Furthermore, it has recently been shown that activation of PPAR␥ by the thiazolidinedione, rosiglitazone, decreases scavenger receptor A expression (16) and increases ABCA1 expression (17). ABCA1 is a member of the ATP-binding cassette transporter family and it is involved in the control of the apoA1-mediated cholesterol efflux from macrophages and other peripheral lipid stores. Mutations in this protein occur in patients with Tangiers disease, a syndrome that is characterized by the pathological accumulation of cholesterol esters in many tissues. Up-regulation of ABCA1 may promote cholesterol clearance, and overall, PPAR␥ appears to be a negative regulator of cholesterol accumulation (16,17). Indeed, in vivo studies using animal models and preliminary clinical data have suggested that PPAR␥ agonists oppose atheroma progression (18 -22).
In contrast to PPAR␣ and -␥, the function of PPAR␦ is relatively unknown. PPAR␦, also known as PPAR␤, NUC1, and FAAR, has been shown to be expressed in a wide range of tissues, but progress in understanding the function of this protein has been hampered by the lack of selective ligands. PPAR␦ has recently been implicated in a wide range of physiological and pathophysiological processes such as embryonic implantation, wound healing, inflammation, cancer, and osteoporosis (23)(24)(25)(26)(27)(28)(29). In this report, we show that PPAR␦ mRNA and protein levels are dramatically elevated during macrophage differentiation and provide both genetic and pharmacological evidence that PPAR␦ is a positive effector of lipid accumulation in human macrophage cultures. We also show that PPAR␦ coordinates a complex pattern of gene expression controlling lipid uptake, transport, storage, metabolism, and efflux.

Isolation of Human Monocytes and in Vitro Monocyte/Macrophage
Differentiation-Human monocytes were isolated from buffy coat preparations of whole blood taken from healthy volunteers. In brief, the buffy coat was mixed with Optiprep TM (Robbins Scientific Ltd.) in a ratio of 2.5:1 and then overlaid with a discontinuous Optiprep TM gradient, prepared according to the reagent datasheet. Following centrifugation for 25 min at 600 ϫ g the monocyte layer formed within the top 5-10 ml of the gradient was removed, washed with phosphate-buffered saline, and resuspended in RPMI 1640, supplemented with 2 mM glutamine and 10% human serum (Sigma). Cell viability was assessed by the ability to exclude trypan blue and was typically 95%. Monocyte purity was determined by differential counts of DiffQuik (Porvair Sciences Ltd.) stained cell preparations and was typically Ͼ95%. For monocyte-macrophage differentiation, monocytes isolated as above were resuspended in culture medium at a density of 2.5 ϫ 10 6 /ml and seeded into 12-well tissue culture plates; medium was changed every 48 h. LDL, oxidized LDL, and lipid-depleted serum was obtained from Intracel Corporation.
Transient Transfection-COS-1 cells were transfected by a DEAEdextran method as previously described (30). The reporter construct pFABPLUC contains 4 copies of the peroxisome proliferator response element from the human liver FABP gene in front of the herpes simplex virus-thymidine kinase promoter, cloned immediately upstream of the cDNA encoding firefly luciferase in pGLBAS (Promega). The PPAR expression vectors contained the coding sequence for human PPAR␣, -␦, and -␥ under the control of the enhancer/promoter of the human cytomegalovirus. PSV␤Gal was co-transfected with each sample to act as an internal control for transfection. Cell lysates were prepared and luciferase and ␤-galactosidase activities were assayed using kits as described by the manufacturer (Promega). Data are presented as the relative light units obtained with the luciferase assay divided by the absorbance obtained at 405 nm in the ␤-galactosidase assays (Luc/␤-galactosidase).
Culture and Differentiation of THP-1 Cells-THP-1 cells were obtained from ATCC. Cultures were grown in RPMI supplemented with 10% heat-inactivated fetal calf serum. Differentiation was initiated by the addition of 5 ng/ml PMA in the above medium. All drugs were added in Me 2 SO and were replaced at intervals of 48 h. After 4 days the cells were fixed with 0.66% paraformaldehyde then stained with Oil red O (Sigma) and hematoxylin (Sigma). Quantification of lipid accumulation was achieved by extracting Oil red O from stained cells with isopropyl alcohol and measuring the optical density of the extract at 510 nm. The value obtained using a control culture was subtracted from the resulting values. The Oil red O absorbance was corrected by co-staining DNA with SYBR green dye (Molecular Probes) and quantified on a Labsystems FluoroSkan Ascent FL microplate fluorimeter. Cell number was determined from a standard curve.
Isolation of Stable Cell Lines-The entire coding sequence of human PPAR␦ was subcloned, in both sense and antisense orientations, into the eukaryotic expression vector pCLDN (31). The resulting plasmids were transfected into THP-1 cells using a modified DEAE-dextran procedure (32). The cells were maintained in medium containing 1 mg/ml G418 and 10% THP-1 conditioned medium, with rigorous washing procedures to remove dead cells. This was continued until cell killing stopped and robust growth was observed. We obtained six independently transfected polyclonal populations each of pCLDN (empty vector) and pCLDNPPAR␦ (SENSE).
Western Blotting-Cultures were lysed in SDS-polyacrylamide gel electrophoresis loading buffer and analyzed by Western blotting using standard procedures. The PPAR␦ antiserum (obtained from Dr. David Bell) was used at a 1:2000 dilution and the anti-AFAR antiserum (obtained from Dr. John Hayes) was used at a 1:3000 dilution. A peroxidase-conjugated mouse anti-rabbit IgG antiserum (Sigma) was used as a secondary detection reagent at 1:3000 and the results were visualized using enhanced chemiluminescence (ECLϩ) as described by the manufacturer (Amersham Pharmacia Biotech).
Lipid Extraction and Measurement-Cultures were treated as above and then extracted with methanol/chloroform/phosphate-buffered saline (1:1:1, v/v) containing stigmasterol as an internal standard for cholesterol. The organic phase was analyzed for cholesterol with and without saponification by GC-MS using a Finnegan ThermoQuest Trace 2000. Triglycerides were quantified colorimetrically using a kit from Sigma and triolein as a standard. Protein concentrations were determined using the Bio-Rad Protein assay reagent.
Cholesterol Efflux Assays-THP-1 cells were differentiated with PMA in the absence or presence of 100 nM compound F for 3 days. PPAR␦ SENSE cells were differentiated with PMA for 3 days. The cells were labeled with 2 Ci (40 nmol) of [4-14 C]cholesterol (Amersham Pharmacia Biotech) in 2 ml of medium containing 10% fetal calf serum for 3 h. ApoA1 specific efflux was determined as previously described (33). For apoA1 independent efflux assays the cells were incubated a further 120 h in medium containing 10% fetal calf serum. Radioactivity release into the medium was determined by scintillation counting at 1, 4, 18, 24, 48, 72, and 120 h. All cultures incorporated similar levels (between 80 and 95%) of the labeled cholesterol. Cell viability was similar between cultures using this procedure.
RNA Extraction and Analysis-RNA was extracted from THP-1 cells using TRIZOL reagent as recommended by the manufacturer. cDNA was synthesized from such RNA using "You-Prime Ready-To-Go" beads from Amersham Pharmacia Biotech. TAQMAN TM real time PCR analysis was applied using prepared reagents from PerkinElmer Life Sciences. The primers and probes will be described elsewhere. 2 Relative levels of mRNA were calculated using the values obtained with each target gene compared with the values obtained with the 18 S ribosomal RNA probes.
Statistical Analysis-All graphs and statistics were prepared using Graphpad Prism for the Macintosh v3.0 (Graphpad Inc., San Diego, CA). Nonlinear regression was applied using a sigmoidal response model for the dose response to compound F in transient transfection experiments (Fig. 3). p values displayed in Figs. 7, B and C, and 8E were calculated using a standard Student's t test.

PPAR␥ Accumulation in Macrophages Is Accompanied by Increases in the Expression of PPAR␦-To examine the role of
PPARs in macrophage lipid accumulation we measured the levels of PPAR protein and mRNA during the differentiation of primary human macrophages in vitro. We confirmed previous observations that PPAR␥ protein levels are increased during this process (5, 6, 8, 10) (Fig. 1A). Interestingly, we also found that PPAR␦ levels are markedly increased during macrophage differentiation. PPAR␣ protein was also detectable, but was not greatly altered during differentiation. This pattern of regulation was confirmed at the mRNA level (Fig. 1B).
PPAR␥ Activation Negatively Regulates Lipid Accumulation in Macrophages-We then examined the effect of rosiglitazone, a highly selective PPAR␥ agonist, on the accumulation of lipid in primary human macrophages under normal culture conditions (10% human serum) (Fig. 2, A versus B). We observed that rosiglitazone (1 M) partially inhibited the accumulation of lipid under such conditions. We also found that rosiglitazone did not stimulate lipid accumulation in primary macrophages in lipid-depleted serum alone, or supplemented with 100 g/ml normal or oxidized low density lipoproteins (Fig. 2C, nLDL and oxLDL, respectively).
Activation of PPAR␦ Promotes Lipid Accumulation in Primary Human Macrophages-In view of the marked induction of PPAR␦ during human macrophage differentiation, we investigated whether it may be a regulatory factor in the process of lipid accumulation. Investigations into the role of PPAR␦ in lipid metabolism have been hampered by the lack of PPAR selective agonists. A series of compounds has recently been described as having selectivity for PPAR␦ (33,34), and we have synthesized one of these compounds (compound F) that has an EC 50 of 2 nM for hPPAR␦, 400 nM for hPPAR␥, and Ͼ1 M for hPPAR␣ (Fig. 3A). Interestingly, this compound has no selectivity between mouse PPAR␥ and -␦ (data not shown). Human macrophages were therefore treated with compound F (100 nM) in the presence of serum, resulting in a profound increase in lipid accumulation (Fig. 3B). The formation of foam cells is thought to be promoted by modified lipoproteins in vivo. We therefore tested the ability of compound F to facilitate lipid accumulation in the presence of oxidized LDL (Fig. 3C). Oxidized LDL is a stimulus for lipid accumulation in human macrophages and addition of increasing concentrations of compound F stimulates further increases in lipid accumulation. The dose response of this lipid accumulation was very similar to the concentrations required for activation of PPAR␦ in COS-1 cells and occurred at concentrations well below those required for the activation of PPAR␥ and -␣. Importantly, lipid loading was not observed with normal LDL or lipid-depleted serum at any concentration of compound F (data not shown). This demonstrates that the action of compound F is dependent on the nature and availability of the lipid donors in the culture medium.
PPAR␦ Promotes Lipid Accumulation in the Monocytic Cell Line THP-1-To obtain further evidence that PPAR␦ is involved in macrophage lipid accumulation we utilized a cell line model where the levels of PPAR␦ could be modulated genetically. The human monocytic cell line, THP-1, can be differentiated into adherent macrophage-like cells by the addition of phorbol ester to the culture medium. Measurement of PPAR mRNA levels following phorbol ester treatment showed that PPAR␣ expression remained unchanged during differentiation (Fig. 4A). The levels of both PPAR␥ and -␦ mRNA were greatly  increased upon differentiation (50-and 34-fold, respectively). These findings are similar to those obtained during the differentiation of primary macrophages (Fig. 1). To determine whether the PPARs were involved in lipid accumulation, THP-1 cells were treated with activators of PPAR␣, -␥, and -␦ in the presence of serum. No change in lipid accumulation was observed on the addition of the PPAR␣ activator, Wy14,643, or the PPAR␥ activator, rosiglitazone (Fig. 4B). However, when THP-1 cells were treated with compound F, a marked increase in intracellular lipid was observed (Fig. 4B). The EC 50 of compound F for this effect was higher than that observed for the primary monocytes (ϳ400 nM, Fig. 4C). The effect of compound F was further enhanced by the addition of the RXR agonist, LG100268 (Fig. 4D); further supporting the action of a PPAR/ RXR heterodimer in this process. There was a high degree of similarity in the findings with those obtained in primary macrophages indicating that THP-1 cells provide a good model for studying the role of PPAR␦ in macrophage lipid accumulation.
PPAR␦ Is a Potent Regulator of Genes Involved in Lipid Accumulation-The lipid accumulation phenotype observed with compound F may reflect regulation of a number of aspects of lipid transport and metabolism in these cells. Recent studies have shown that rosiglitazone modulates lipid uptake by transcriptional regulation of the class B scavenger receptor CD36, but that this is balanced by increases in efflux via ABCA1 and decreases in the class A scavenger receptor, SRA (16). We therefore investigated the effects of compound F on the expression of a range of genes known to determine lipid balance in macrophages. We performed a dose-response experiment with PPAR␦ selective concentrations of compound F and compared these treatments to a saturating dose of rosiglitazone (500 nM) (Fig. 5A). We observed that CD36 is effectively induced (6-fold) by 10 nM compound F, and that this induction is greater than that seen with 500 nM rosiglitazone (2-3-fold). This suggests that CD36 is a target gene for both PPAR␦ and PPAR␥. The class A scavenger receptor was also slightly increased by 10 nM compound F (2.5-fold), whereas rosiglitazone does not modify the expression of this gene. The gene encoding the adipose-type fatty acid-binding protein (AFABP) is highly regulated by both rosiglitazone and 10 nM compound F, again showing an overlapping set of target genes. However, the gene encoding the lipid-vesicle coat protein, adipophilin is very selectively regulated by compound F (10-fold versus 3-fold with rosiglitazone). The selectivity for activation of SRA, CD36, and adipophilin was confirmed in a time course study (Fig. 5B) where it was clear that compound F mediates a larger and more sustained induction of these gene products relative to rosiglitazone. The induction of SRA was very modest but sustained, in contrast no significant induction with rosiglitazone was seen at any time point. CD36 induction by compound F appeared to be diminished at longer time points but adipophilin was highly induced throughout the experiment.
Overexpression of PPAR␦ Results in Increased Lipid Accumulation-To obtain evidence for a direct role for PPAR␦ in macrophage lipid accumulation, we generated THP-1 sublines that overexpressed PPAR␦ (PPAR␦SENSE). Constitutive expression of PPAR␦ in the PPAR␦SENSE cells was confirmed by Western blotting (Fig. 6A). In agreement with the effects of the PPAR␦ agonist, PPAR␦ SENSE cell lines were extremely lipidladen when compared with the wild type cells, even in the absence of added PPAR ligand (but in the presence of serum)  (Fig. 6B). This response is very rapid with very obviously lipidladen cells appearing 3 days after differentiation, whereas compound F promotes lipid accumulation much more slowly, only becoming visible after 7 to 10 days of treatment (data not shown). We used gas chromatography/mass spectrophotometry to measure the levels of free and esterified cholesterol present in THP-1 cells differentiated with PMA with and without compound F (100 nM), and in differentiated PPAR␦SENSE cells (Fig. 6C). We found that compound F treatment resulted in a 50% increase in cholesterol esters compared with the untreated cells, however, the levels of free cholesterol were seen to be increased in the PPAR␦SENSE cells. This difference may lie in the kinetics of lipid accumulation in the SENSE cells where visible lipid accumulation occurs much more rapidly than in the compound F-treated THP-1 cells. This may deplete the cells of cholesterol acceptors within the cell thus resulting in an accumulation of free cholesterol. This hypothesis is supported by the observation that triglyceride levels are massively increased in the PPAR␦SENSE cells (Fig. 6D). Interestingly, triglyceride accumulation in conjunction with cholesterol is a characteristic of human rather than rodent macrophages (35).
Overexpression of PPAR␦ Modulates Expression of Genes Involved in Lipid Transport and Storage-The nonstimulated PPAR␦SENSE cells contain high constitutive levels of AFABP and adipophilin, with AFABP present at levels over 100 times those observed in the untransfected cells (Fig. 6, E and F). This suggests that the nonstimulated PPAR␦SENSE cells have a certain degree of signaling through the overexpressed PPAR␦ protein and provides further confirmation of the role for PPAR␦ in the regulation of these genes. These two genes are the most sensitive to compound F (Fig. 5) and it was notable that none of the other target genes (including CD36) were elevated in nonstimulated PPAR␦SENSE cells (data not shown). This reinforces the concept of a hierarchy of gene targets in PPAR␦ biology. Adipophilin is a marker of lipid loading in macrophages (36) and forms a major component of the coating for the intracellular lipid vesicles. Large amounts of this protein would therefore be required to form the vesicles seen in Fig. 6B (37). In contrast to adipocytes, perilipin is barely expressed in THP-1 cells (data not shown). Perilipin is the final coating for lipid storage in adipocytes and such perilipin-bound vesicles are susceptible to mobilization by PKA. Therefore, it is possible that PPAR␦ may induce a lipid storage pool that is less susceptible to lipolytic signals.
PPAR␦ Represses a Subset of Genes Involved in Cholesterol Efflux-Upon activation of THP-1 cells we found that cholesterol 27-hydroxylase (CYP27) and apolipoprotein E (apoE) were repressed by compound F (Fig. 7, A and C). This repression was also seen in the activated and lipid-filled PPAR␦SENSE cells in the absence of compound F (Fig. 7, B and D). ApoE and CYP27 are important mediators of lipid efflux from macrophages (38), and reduced expression may contribute to the defect in lipid clearance observed in the PPAR␦ overexpressing cells. However, during the preparation of this manuscript there has been a report of the up-regulation of macrophage ABCA1 by the PPAR␦ agonist, GW501516, and a subsequent increase in apoA1-mediated efflux from cells treated with this compound (33). Our results confirm that the expression of ABCA1 is increased by agonists of PPAR␦ (Fig. 5A), however, the finding that apoE and CYP27 gene expression is greatly diminished suggests that total cholesterol efflux may not be increased by these compounds. We therefore compared the effects of compound F and PPAR␦ overexpression on the apoA1-dependent cholesterol efflux and total efflux of radioactive cholesterol into medium containing normal serum. In these studies we found that apoA1-specific efflux was increased by compound F treat-ment and PPAR␦ overexpression (Fig. 7E) in agreement with our observation that ABCA1 is up-regulated and with the study of Oliver et al. (33). In contrast, total efflux from the cells was significantly attenuated by compound F treatment or PPAR␦ overexpression (Fig. 7F). This data supports our gene expression data in suggesting that PPAR␦ may have opposing effects on different lipid efflux pathways. The rank order of effluxinhibition correlated well with the rate of lipid accumulation i.e. SENSE Ͼ compound F Ͼ control THP-1. DISCUSSION We have demonstrated that the expression of PPAR␦ is increased dramatically during the differentiation of human primary macrophages and upon differentiation of THP-1 cells with phorbol ester. This has suggested a potential role for PPAR␦ in the modulation of atherosclerosis and inflammation. The awareness of coincident expression of PPAR␥ and PPAR␦ is important in the interpretation of studies exploring macrophage biology that have used poorly selective PPAR agonists (5,6,10,39). In the current study we have used highly selective agonists to demonstrate that both PPAR␦ and PPAR␥ differentially regulate the lipid accumulation in macrophages, with PPAR␦ activation promoting lipid accumulation and PPAR␥ promoting lipid clearance. These findings are in contrast with studies that suggested PPAR␥ may promote macrophage lipid accumulation; however, these studies utilized compounds that are not as selective for PPAR␥ as rosiglitazone (5,6). Indeed, several publications, which were published during the preparation of this manuscript (7,16,17), support our finding that PPAR␥ activation by rosiglitazone opposes lipid accumulation in macrophages. This role for PPAR␥ in lipid clearance is in distinct contrast to its role in adipocyte differentiation, but is similar to the proposed role for PPAR␥ in the clearance of lipid from pancreatic ␤ cells (40). We have found that PPAR␦ selectively promotes the accumulation of lipid in both THP-1 cells and ex vivo human macrophage cultures. This lipid accumulation occurs because of a complex regulation of gene products that control many aspects of lipid homeostasis (Fig. 8). PPAR␦ controls lipid uptake, via the class A and class B scavenger receptors (SR-A, CD36). Transport is mediated by increases in lipid-binding proteins such as AFABP. Storage is facilitated by the production of large amounts of lipid vesicle coating proteins such as adipophilin, while metabolism of cholesterol to bile acids is regulated by the repression of CYP27. Lipid efflux is modulated in a contrasting manner by the induction of ABCA1 and the repression of apolipoprotein E. It is clear that this list of PPAR␦-regulated genes is incomplete and that the dramatic lipid storage phenotype that we have observed is determined by a complex interplay of such target genes. Importantly, we have characterized the novel repression of two gene products that are intimately involved in lipid storage disorders, i.e. apoE and CYP27. It is clear that both of these proteins are important in human lipid homeostasis. Indeed, genetic studies have shown that functional polymorphisms of the human apoE gene are associated with atherosclerosis (41)(42)(43)(44) and mutations in the CYP27 gene are seen in patients with cerebrotendinous xanthomatosis (45). This is a severe cholesterol storage disorder where cholesterol is deposited throughout the body, but is primarily targeted to the vasculature and brain. The metabolism of cholesterol by this enzyme in the macrophage directs cholesterol to bile acid synthesis and export, independent of apoA1 lipid acceptors (46 -49). Therefore the profound repression of CYP27 and apoE expression by activation or overexpression of PPAR␦ might contribute to the lipid accumulation observed in these cells in culture. Importantly, although PPAR␦ and PPAR␥ have some common target genes, apoE and CYP27 are not repressed by PPAR␥ agonists such as rosiglitazone. PPAR␥ agonists have been consistently shown to reduce the atherogenic profile of serum lipids (50,51) and have inhibitory effects on atheroma formation (18 -22). We and others have shown that this is reflected in the modest ability of rosiglitazone to promote lipid clearance in human monocytes in vitro and it is likely that the anti-inflammatory nature of PPAR␥ agonists also contribute to inhibition of atherosclerotic processes. In contrast, PPAR␦ agonists have been shown to have complex effects on serum lipids in different animal models, with a common feature of raising total serum cholesterol (33,50,52). The relevance to man of studies of PPARs on lipids in small animals is often limited. One of these studies, using the highly selective PPAR␦ agonist, GW501516, has also shown a potent raising of serum high density lipoprotein and lowering of serum triglycerides in obese rhesus monkeys (33). The mechanism of triglyceride lowering by GW501516 was distinct from the mechanism of lipid lowering seen in the use of PPAR␣ ligands such as bezafibrate. The repression of apoC-III expression by fibrates is considered to be an important factor in lipid lowering and bezafibrate treatment of obese rhesus monkeys produces marked decreases in serum apoC-III, whereas GW501516 produced a marked increase in serum apoC-III. GW501516 has no effects on apoC-III gene expression in cultured human hepatocytes and the mechanism of lipid lowering by GW501516 is yet to be determined. It is therefore important that future studies address the effects that PPAR␦ agonists have on both the processing of lipid in specific lipid depots, and whole body lipid balance. However, it is clear that PPAR␦ agonists behave differently in different animal models and such studies will have to be chosen with great care. Our study has focused on the pharmacology and molecular biology of the human receptor in human cells and we have demonstrated that PPAR␦ mediates a complex program of gene expression for the FIG. 7. Cholesterol efflux is modulated selectively by PPAR␦. PMA-differentiated THP-1 cells were treated with increasing concentrations of compound F (CF) or 500 nM rosiglitazone (Ros) for 48 h and analyzed for cholesterol 27-hydroxylase (A) and apolipoprotein E (C) mRNA levels by TAQMAN TM procedures. PMAdifferentiated THP-1 and PPAR␦SENSE cells were analyzed for cholesterol 27-hydroxylase (B) and apolipoprotein E (D) mRNA levels by TAQMAN TM procedures. E, THP-1 cells were differentiated with PMA in the absence or presence of 100 nM compound F for 3 days. PPAR␦SENSE cells were differentiated with PMA for 3 days. The cells were labeled with 14 C cholesterol and apoA1-specific efflux was determined. Shown is mean Ϯ S.E., n ϭ 3. F, cells were labeled with [ 14 C]cholesterol as above. Efflux was measured in the presence of serum between 18 and 120 h. No significant differences were observed between 1 and 18 h (data not shown). Results shown are the mean Ϯ S.E., n ϭ 6. DMSO, dimethyl sulfoxide.
FIG. 8. PPAR␦ modulates many aspects of lipid homeostasis in macrophages. Shown is a schematic of various processes, which are controlled by PPAR␦, and determine the lipid storage phenotype. control of lipid accumulation in human macrophages. Further investigations will be required to determine the usefulness of PPAR␦ agonists and antagonists in the management of human metabolic and vascular disease.