Activation of Proliferator-activated Receptors α and γ Induces Apoptosis of Human Monocyte-derived Macrophages*

Peroxisome proliferator-activated receptors (PPARs) have been implicated in metabolic diseases, such as obesity, diabetes, and atherosclerosis, due to their activity in liver and adipose tissue on genes involved in lipid and glucose homeostasis. Here, we show that the PPARα and PPARγ forms are expressed in differentiated human monocyte-derived macrophages, which participate in inflammation control and atherosclerotic plaque formation. Whereas PPARα is already present in undifferentiated monocytes, PPARγ expression is induced upon differentiation into macrophages. Immunocytochemistry analysis demonstrates that PPARα resides constitutively in the cytoplasm, whereas PPARγ is predominantly nuclear localized. Transient transfection experiments indicate that PPARα and PPARγ are transcriptionally active after ligand stimulation. Ligand activation of PPARγ, but not of PPARα, results in apoptosis induction of unactivated differentiated macrophages as measured by the TUNEL assay and the appearance of the active proteolytic subunits of the cell death protease caspase-3. However, both PPARα and PPARγ ligands induce apoptosis of macrophages activated with tumor necrosis factor α/interferon γ. Finally, PPARγ inhibits the transcriptional activity of the NFκB p65/RelA subunit, suggesting that PPAR activators induce macrophage apoptosis by negatively interfering with the anti-apoptotic NFκB signaling pathway. These data demonstrate a novel function of PPAR in human macrophages with likely consequences in inflammation and atherosclerosis.

Peroxisome proliferator-activated receptors (PPARs) 1 are transcription factors belonging to the nuclear receptor gene family (1,2). PPARs are characterized by distinct tissue distribution patterns and metabolic functions. PPAR␣ is predomi-nantly expressed in tissues exhibiting high catabolic rates of fatty acids (liver, heart, kidney, and muscle) (3,4), while PPAR␥ is adipose tissue selective where it triggers adipocyte differentiation and lipid storage by regulating the expression of genes critical for adipogenesis (5,6). PPARs function as liganddependent transcription factors, which, upon heterodimerization with the 9-cis-retinoic acid receptor, bind to specific response elements termed peroxisome proliferator-response element (PPRE), thus regulating the expression of target genes. Most PPREs identified to date reside in genes involved in intraand extracellular lipid metabolism (for review, see Ref. 2).
The natural prostaglandin 15-deoxy-⌬ 12,14 -prostaglandin J 2 (PG-J2) (7,8) and the synthetic antidiabetic thiazolidinediones (9) are ligands for PPAR␥, while hypolipidemic fibrates and eicosanoids, such as leukotriene B 4 (10) and 8(S)-hydroxyeicosatetraenoic acid (11,12), are synthetic and natural ligands, respectively, for PPAR␣. The observation that PPARs are activated by arachidonic acid metabolites suggested a role for these transcription factors not only in lipid metabolism, but also in inflammation control. Furthermore, both PPAR␣ and PPAR␥ are activated by a number of non-steroidal anti-inflammatory drugs, such as indomethacin (13). Unequivocal evidence for a role of PPAR␣ in inflammation control came from the demonstration that mice rendered deficient for PPAR␣ by homologous recombination display a prolonged response to inflammatory stimuli (10).
Formation of atherosclerotic lesions is a multicellular process in which lipids and extracellular matrix accumulate in the intima of arteries and a local inflammatory response is elicited due to the activation of macrophages, T-lymphocytes, smooth muscle, and endothelial cells (14 -16). Lipid-loaden foam cells, derived principally from monocyte-derived macrophages (17)(18)(19), are a characteristic feature of atherosclerotic plaques (20). The accumulation of macrophages in the arterial wall follows adhesion of circulating monocytes to the endothelium and their subsequent transmigration into the arterial intima (17). In the subendothelial space monocytes mature into tissue macrophages and acquire the ability to recognize oxidatively modified lipoproteins by expressing scavenger receptors (21,22). When macrophages take up these lipoproteins, the cholesterol is stored in the cytoplasm in cholesteryl ester droplets, which give the cells its foamy appearance, thus accounting for the term foam cells (23). Intracellular accumulation of cholesterol triggers macrophages to secrete cytokines, growth factors, and other mediators that promote smooth muscle cell proliferation and provoke a local inflammatory reaction (15,24). Apoptosis, a form of programmed cell death involved in tissue morphogenesis and homeostasis (25), may be induced by these cytokines resulting in cell death of macrophages and smooth muscle cells in atherosclerotic lesions (26). In addition to apoptosis, in advanced atherosclerotic plaques, macrophages and lipid-loaded foam cells undergo necrosis thereby releasing their intracellular contents due to cytolysis resulting in the formation of the lipid-rich core of the atheromatous plaque (27).
Considering the role of PPARs in lipid metabolism and inflammatory control, we hypothesized a function for PPARs at the level of the vascular wall which, independent of their role in lipoprotein metabolism, could modulate the pathogenesis of atherosclerosis. Therefore, we initiated studies on the expression and functions of PPARs in circulating human monocytes and in macrophages during differentiation, a crucial event in atherosclerosis development and vascular inflammation. Our results show that PPAR␣ is already expressed in undifferentiated human monocytes, whereas PPAR␥ only becomes expressed in human macrophages upon differentiation. Treatment of differentiated macrophages with PPAR activators induces apoptosis, an effect which is more pronounced in macrophages activated with IFN␥ and TNF␣. Finally, we show that PPAR␥ inhibits the transcriptional activity of the NFB subunit p65/RelA, indicating that PPAR␥ activators may promote TNF␣-induced apoptosis in macrophages by interfering negatively with the anti-apoptotic NFB pathway (28 -30).

EXPERIMENTAL PROCEDURES
Cell Culture-Mononuclear cells were isolated from blood of healthy normolipidemic donors (thrombopheresis residues) (31). After Ficoll gradient centrifugation, the monocytes were suspended in RPMI 1640 medium containing gentamycin (40 g/ml), glutamine (0,05%) (Sigma), and either 5% pooled human serum and 5% fetal calf serum (apoptosis assays) or 10% pooled human serum (other assays). Cells were cultured at a density of 3 ϫ 10 6 cells/well in 6-well plastic culture dishes (Primaria, Polylabo, Strasbourg, France). Differentiation of monocytes into macrophages occurred spontaneously by adhesion of cells to the culture dish. Mature monocyte-derived macrophages, as characterized by immunocytochemistry with anti-CD 68 antibody, were used for experiments after 12 days of culture. For treatment with the different activators medium was changed to RPMI 1640 medium with 1% Nutridoma HU (Boehringer Mannheim).
For RNase protection analysis, an antisense human PPAR␣ riboprobe was obtained by subcloning a SmaI/SacI human PPAR␣ cDNA fragment (spanning bp 800 -983 (32)) into the pBSKS vector (Stratagene, La Jolla, CA). PPAR␣, covering 54 bp of vector sequence plus 183 bp of PPAR␣ mRNA, and 36B4 riboprobes were in vitro transcribed in the presence of [ 32 P]CTP (800 Ci/mmol) using T3 RNA Polymerase and the RNA transcription kit (Stratagene). The RNA protection assay was carried out using the HybSpeed RNase protection kit (Ambion, Austin, TX). Total RNA (10 g) was hybridized simultaneously to human PPAR␣ (8.4 ϫ 10 4 cpm) and 36B4 (4 ϫ 10 3 cpm) antisense probes. RNA samples were electrophoresed on a 5% denaturing polyacrylamide gel and visualized by autoradiography. Protected fragments representing human PPAR␣ and 36B4 mRNA were quantified on a GS525 Phosphor-Imager (Bio-Rad). Values were normalized to the internal 36B4 control.
Protein Extraction and Western Blot Analysis-Cells were washed twice in ice-cold phosphate buffer saline (PBS) and harvested in ice-cold lysis buffer containing PBS, 1% Triton X-100, and a freshly prepared protease inhibitor mixture (ICN, Orsay, France) (10 g/ml AEBSF, 1 mg/ml leupeptin, 1 mg/ml pepstatin, 5 mg/ml EDTA-Na 2 ) to which 1 mM phenylmethylsulfonyl fluoride was added. Cell homogenates were col-lected by centrifugation at 13,000 rpm at 4°C and protein concentrations were determined using the bicinchonic acid assay (Pierce Interchim, Montlucon, France). Electrophoresis of the indicated amount of protein lysate was performed through a 10% polyacrylamide gel under reducing conditions (sample buffer containing 10 mM dithiothreitol). Proteins were transferred onto nitrocellulose membranes and membranes were checked for equal loading by Ponceau red staining. Nonspecific binding sites were blocked overnight at 4°C with 10% skim milk powder in TBST (20 mM Tris, 55 mM NaCl, 0.1% Tween 20). Membranes were subsequently incubated for 4 h at room temperature in 5% skim milk-TBST containing rabbit polyclonal antibodies raised against N-terminal PPAR␣ (amino acids 10 -56), PPAR␥ peptides (33), or a monoclonal antibody against caspase-3 (Transduction Laboratories, Montlucon, France). Specificity of PPAR antibodies was checked by Western blot analysis using in vitro synthesized PPAR␣ and PPAR␥ protein (Fig. 2B). After incubation with a secondary peroxidase-conjugated antibody, signals were visualized by chemiluminescence (Amersham, Buckinghamshire, United Kingdom).
Transient Transfections-The J 3 TKpGL3 plasmid, which contains 3 copies of the human apoA-II gene promoter PPRE-containing J site cloned upstream of the thymidine kinase (TK) promoter in the pGL3 luciferase expression vector (34), was used as a reporter vector. Monocyte-derived macrophages at day 12 of culture were transfected by the lipofection procedure (DOTAP; Boehringer Mannheim, Mannheim, Germany) with the reporter plasmid (2.5 g/well of a 6-well culture dish) in RPMI 1640 medium containing 1% Nutridoma HU. After 5 h at 37°C, cells were washed with PBS and ligands for PPAR␣ (Wy14, 643; Chemsyn, Lenexa, KS), PPAR␥ (BRL49653 and PG-J2), or solvent dimethyl sulfoxide were added in RPMI 1640 medium containing 5% calf serum delipoproteinized by ultracentrifugation in KBr (1.21 g/ml) and subsequently treated with AG.1.X8 resin (Bio-Rad) plus activated charcoal, and cells were incubated further for 36 h.
For NFB and PPAR␥ co-transfection assays, transient transfections were performed in COS-1 cells at 50 -60% confluency by the calciumphosphate co-precipitation procedure using a mixture of plasmids containing the indicated reporter (2 g/60-mm culture dish), expression (1 g) and internal control pPGK␤geobpA (500 ng) ␤-galactosidase activity (35) vectors. After 6 h cells were washed with PBS and incubated in fresh medium containing 0.5% fetal calf serum with BRL49653 (10 M) or vehicle (Me 2 SO) and incubated for another 48 h. Luciferase activity was determined on cell extracts using a luciferase buffer (Promega, Madison, WI).
Detection Of Apoptotic Cells-Monocyte-derived macrophages (12 days of culture) were incubated for 24 h at 37°C in RPMI 1640 medium with 1% Nutridoma HU, TNF␣, and IFN␥ and/or Wy14,643, BRL49653, PG-J2, or solvent (Me 2 SO). Apoptotic cells were detected using the In Situ Cell Death Detection and Cell Death ELISA kits (Boehringer Mannheim).

PPAR␣ and PPAR␥ Are Expressed in Differentiated Human
Monocyte-derived Macrophages-To study the expression of PPAR␣ and PPAR␥ in human monocyte-derived macrophages, a qualitative RT-PCR analysis was performed using specific primers on RNA from freshly isolated human monocytes and macrophages at different stages of differentiation. RT-PCR amplification yielded a DNA fragment of the expected size (920 bp) indicating the presence of PPAR␣ mRNA both in undifferentiated monocytes as well as in macrophages during all stages of differentiation (Fig. 1A). Using a quantitative RNase protection assay human PPAR␣ mRNA was detected in monocytes albeit at low levels (Fig. 1B). PPAR␣ mRNA levels increased between day 0 (monocytes) and day 4 (cells undergoing differ-entiation), staying constant during the following maturation phases (days 6 -10) and further increasing in fully differentiated macrophages after 12 days of culture (Fig. 1B). In contrast to PPAR␣, PPAR␥ mRNA was not detectable by RT-PCR analysis in non-differentiated cells (Fig. 1A). However, PPAR␥ mRNA was induced during the first stages of macrophage differentiation staying present thereafter (Fig. 1A). Similar results of PPAR␥ expression were obtained by Northern blot analysis (data not shown).
To determine PPAR protein expression in monocytes and differentiated macrophages, Western blot analysis was performed using PPAR␣-and PPAR␥-specific antibodies, as demonstrated by Western blot analysis using in vitro produced PPAR␣ and PPAR␥ protein (Fig. 2B). A single band corresponding to PPAR␣ was detected in undifferentiated monocytes ( Fig.  2A). Furthermore, the amount of PPAR␣ protein increased upon differentiation (Fig. 2A). By contrast, PPAR␥ protein was not detectable in monocytes, but its expression appeared in cells undergoing differentiation into macrophages ( Fig. 2A). Therefore PPAR␣ and PPAR␥ protein levels closely follow the changes in mRNA levels ( Figs. 1 and 2).
Cytolocalization of PPAR␣ and PPAR␥-To identify the subcellular localization of PPAR␣ and PPAR␥ in differentiating macrophages, immunofluorescence analysis was performed. Staining with PPAR␣ antibody revealed immunoreactivity in monocytes as well as in macrophages during all phases of the differentiation process (Fig. 3, A, E, I, and O). Interestingly, labeling was detected primarily in the cytoplasmic compartment, indicating that PPAR␣ protein resides predominantly in the cytoplasm of macrophages (Fig. 3, O). Consistent with the results from the Western blot experiments ( Fig. 2A), no label-ing with anti-PPAR␥ antibody was observed in monocytes (Fig.  3B). However, in differentiating cells, immunoreactive staining was observed in the nucleus (Fig. 3, L and P), indicating that PPAR␥ is predominantly nuclear localized in macrophages.
As a control for the monocyte-macrophage maturation process, monocytes and lymphocytes were identified using specific marker antibodies for macrophages (CD68) or lymphocytes (CD3). Along the differentiation process, cells became progressively reactive for the CD68 antibody and after 12 days of culture all adherent cells were CD68 positive (Fig. 3, D, H, N,  and R). In contrast, whereas at day 0 a number of cells stained positive for CD3, this staining decreased progressively throughout differentiation disappearing completely at day 12, thus indicating the absence of T cells after 12 days of culture ( Fig. 3, C, G, M, and Q).
Endogenous PPARs Can be Ligand Activated-Since PPARs are transcription factors, it was determined whether endogenous PPAR␣ and PPAR␥ are transcriptionally active in differentiated macrophages. Therefore, macrophages were transiently transfected with the PPRE-driven J 3 TKpGL3 luciferase reporter vector and cells were subsequently treated with PPAR␣-and PPAR␥-specific ligands. Treatment with the PPAR␣ ligand Wy14,643 at a concentration (10 M) selectively activating PPAR␣, but not PPAR␥, as measured by a co-transfection transactivation assay (data not shown), resulted in a 2-fold induction of luciferase activity (Fig. 4). Treatment with either BRL49653 or PG-J2 resulted in a stronger induction of luciferase activity which was dose-dependent (Fig. 4).
Induction of Macrophage Apoptosis by PPAR Activators-Interestingly, treatment of differentiated macrophages with BRL49653 or PG-J2 at concentrations within the range of their K D for PPAR␥ binding consistently induced pronounced changes of cellular morphology resulting eventually in cell death. Since these effects were very marked, were reproduced in several experiments using different macrophage preparations, and occurred primarily with the PPAR␥ ligands BRL49653 and PG-J2 at concentrations below those classically used in transactivation and adipocyte differentiation assays (and much less with the PPAR␣ ligand Wy14,643 at 100-fold higher concentrations), it was analyzed whether PPAR activation of macrophages induced cellular apoptosis, a form of programmed cell death. Therefore, the induction of apoptosis was measured by the TUNEL assay in differentiated human monocyte-derived macrophages treated with either Wy14,643, BRL49653, or PG-J2 at a concentration and time point before changes in cell morphology became microscopically apparent. Both BRL49653 and PG-J2 induced an intense nuclear staining (Fig. 5, C and D), indicating the appearance of apoptotic cells. The number of cells staining positively after PG-J2 treatment increased in a dose-dependent fashion (data not shown), similar as the activation of PPAR␥ transcriptional activity (Fig.  4). In contrast, when cells were treated with Wy14,643 at concentrations specifically activating PPAR␣, staining was comparable to solvent background (Fig. 5, A and B), indicating absence of apoptosis induction by Wy14,643. Next, it was analyzed whether PPAR activators could promote apoptosis in the presence of TNF␣, a known inducer of macrophage apoptosis. Activation of differentiated macrophages by TNF␣ and IFN␥ resulted in a significant number of cells staining positive by TUNEL labeling (Fig. 5E). When macrophages were simultaneously treated with TNF␣/IFN␥ and BRL49653 or PG-J2, nuclear labeling was observed in a high number of cells which stained much more intensely than in the presence of each compound alone (Fig. 5, G and H). Interestingly, in the presence of TNF␣/IFN␥, treatment with Wy14,643 resulted also in a significant number of cells stained positive by TUNEL labeling (Fig. 5F). Quantitative analysis of apoptosis showed that treatment with BRL49653 resulted in an approximate 2-fold induction of cellular DNA fragmentation compared with solvent, whereas Wy14,643 was without effect (Fig. 6). Activation of macrophages with TNF␣/IFN␥ resulted in a 2-fold induction of cellular DNA fragmentation, an effect which was enhanced in cells simultaneously treated with TNF␣/IFN␥ and either BRL49653 or Wy14,643 (Fig. 6).
Since cells undergoing apoptosis execute the death program by activating cysteine proteases of the caspase family (36), it was analyzed whether treatment of differentiated macrophages with PPAR activators resulted in the activation of caspase-3/cystein protease protein 32, a key executioner of apoptosis (37). Western blot analysis showed the appearance of the enzymatically active 17-and 12-kDa proteolytic cleavage products of the inactive 32-kDa caspase-3 precursor in differentiated macrophages treated for 24 h with either BRL49653, PG-J2, or TNF␣/IFN␥ alone (Fig. 7). Interestingly, compared with each compound alone, simultaneous activation of cells with TNF␣/IFN␥ and BRL49653 or PG-J2 significantly increased the amount of 17-and 12-kDa cleavage products (Fig.  7). In the presence of both TNF␣/IFN␥ and Wy14,643 a similar but less pronounced increase was observed (Fig. 7). Altogether these data indicate that PPAR activators potentiate TNF␣induced apoptosis.
Since during the preparation of this article it was shown that PPAR␥ activators can negatively regulate transcription from a NFB-response element driven promoter (38) and since NFB via its p65/RelA subunit has been shown to protect macrophages from TNF␣-induced cell death (28 -30), the effects of PPAR␥ on the transcriptional activity of the p65/RelA subunit of NFB was tested by transient co-transfection assays in COS-1 cells. Co-transfection of p65/RelA resulted in a more than 600-fold induction of luciferase reporter activity driven by a NFB response element (Fig. 8A). This induction by p65/RelA was almost completely inhibited by co-transfection of PPAR␥, an effect which was further enhanced in the presence of BRL49653 (Fig. 8A). By contrast PPAR␥ alone did not activate the NFB response element driven reporter vector (Fig. 8A), whereas transcription from a PPRE driven promoter was significantly enhanced by PPAR␥ co-transfection (Fig. 8B). Altogether, these data indicate that PPAR activators induce cellular death of macrophages by apoptosis, most likely by negatively interfering with the antiapoptotic NFB signaling pathway.

DISCUSSION
PPARs are lipid-activated transcription factors that function as important regulators of lipid and glucose metabolism, adipocyte differentiation, and energy homeostasis. PPAR␣ and PPAR␥ mediate, respectively, the action of the hypolipidemic fibrates and the anti-diabetic thiazolidinediones. PPARs therefore play a role in metabolic conditions, such as dyslipidemia and type II diabetes, leading to atherosclerosis development. Several indirect observations suggest that PPAR activators not only regulate plasma cholesterol and triglyceride concentrations, major risk factors for atherosclerosis, but may also exert an activity at the level of the vascular wall. First, in cholesterol-fed rabbits, treatment with the PPAR␣ activator fenofibrate decreases atherosclerotic plaque formation in thoracic aorta, in the absence of any effect on lipoprotein levels (39), suggesting a direct anti-atheromatous action of fibrates. Second, in the BECAIT (Bezafibrate Coronary Atherosclerosis Intervention Trial) (40) and LOCAT (Lopid Coronary Angiography Trial) (41) intervention trials, fibrate treatment prevented the progression of coronary atherosclerosis, in the absence of a marked lowering of plasma atherogenic lipoprotein concentrations. Third, treatment with the PPAR␥ activator troglitazone inhibits vascular smooth muscle cell proliferation (42) and decreases the intima and media thickness of carotid arteries in man (43). Finally, although no in vivo data are available for PPAR␥, PPAR␣ has been shown to play a role in anti-inflammatory control, since PPAR␣ knockout mice exhibit a prolonged inflammatory response (10). Therefore, PPAR activators may not only interfere with atherogenesis by decreasing plasma lipid concentrations, but also by exerting anti-inflammatory activity at the vascular wall.
Macrophages are key players in vascular wall inflammation and atherogenesis. In this report, we demonstrate the expression of both PPAR␣ and PPAR␥ in differentiated human macrophages. Furthermore, we show that PPAR␣ expression is already detectable in monocytes and increases along the differentiation process into macrophages. By contrast, PPAR␥ expression is not detectable in monocytes, but is strongly induced upon differentiation into macrophages. Recently it was reported that PPAR␥ may promote monocyte/macrophage differentiation (44). Furthermore, PPAR␥ is activated by oxidized lipid components present in atherogenic oxidized low density lipoprotein and PPAR␥ activation results in the induction of macrophage scavenger receptor CD36 expression and enhanced uptake of oxidized low density lipoprotein (45). These observations may point to a role for PPAR␥ in foam cell formation (45). Since PPAR␥ expression is undetectable in circulating human monocytes and appears only several hours after induction of differentiation, it is unlikely that PPAR␥ is involved in the initial differentiation process. Furthermore, in these studies maximal activity on monocyte/macrophages was observed with the less specific PPAR␥ agonist PG-J2 in the presence of 9-cis-retinoic acid receptor agonists (44,45), suggesting that activation of additional signaling pathways is required to stimulate these differentiation processes. Nevertheless, PPAR␥ appears to be an important component of further downstream processes in macrophage differentiation and function.
Our transient transfection experiments demonstrate that endogenous PPARs are transcriptionally activated by their ligands at K D concentrations confirming a functional role for PPARs in transcription control of human macrophages. Whereas BRL49653 and PG-J2 stimulate gene transcription severalfold, Wy14,643 appears less active. This low activation of PPAR␣ by Wy14,643 might be due to either the small amount of PPAR␣ in the nucleus compared with PPAR␥, or to the relative weaker affinity of Wy14,643 to human PPAR␣ than BRL49653 to PPAR␥ (9,10). The observation that PPARs are transcriptionally active raises the question on which are the target genes for PPARs in human macrophages. Since PPARs regulate lipid and lipoprotein metabolism, genes involved in macrophage lipid metabolism, such as lipoprotein lipase, which is expressed in macrophages and foam cells (44,45), are likely candidates. Interestingly, a PPRE was identified in the human lipoprotein lipase gene promoter, via which fibrates and thiazolidinediones regulate lipoprotein lipase expression in liver and adipose tissue, respectively (46). Furthermore, the macrophage scavenger receptor CD36, which also functions as a fatty acid transporter in adipose tissue (47), has been identified as a PPAR␥ target gene in macrophages (44). In addition, PPARs also regulate genes involved in inflammatory control, since PPAR␥ activators have been shown to inhibit the activation of macrophages by interfering with the transcriptional induction of genes such as TNF␣, iNOS, and gelatinase B by inflammatory agents (38,48).
In the present report, we demonstrate a role for PPARs in the control of macrophage apoptosis. PPAR␥ activation results in a pronounced induction of apoptosis, whereas PPAR␣ activators are much less pro-apoptotic in unstimulated differentiated human macrophages. This induction of apoptosis by PPAR␥ activators occurs at concentrations well within the range of their K D values. By contrast, the recently reported anti-inflammatory activity of synthetic PPAR␥ activators are observed only at much higher concentrations (10 2 -10 3 -fold above the K D ) (38). Similarly, induction of scavenger receptor CD36 and foam cell formation occurs at severalfold higher concentrations (44). Therefore, the induction of apoptosis might be the physiologically primary event, resulting in a secondary general inhibition of macrophage activation. Alternatively, PPAR␥ activators may have a dual activity: at low concentrations they might be proaptotic and, as a result, prevent foam cell formation, whereas at high concentrations they may favor foam cell formation through induction of scavenger receptors in those macrophages which rescue from apoptosis.
Apoptosis induction by PPAR␥ ligands was more pronounced in macrophages activated with TNF␣ and IFN␥. Furthermore, PPAR␣ ligands induced apoptosis only in activated macrophages. In macrophages, TNF␣ affects cell survival via two distinct signaling pathways with opposing functions: a proapoptotic NFB-independent and an anti-apoptotic NFB-dependent pathway (49). Indeed, in macrophages NFB activation by TNF␣ induces a negative feedback mechanism protecting the cells from TNF␣-induced apoptotic cell death (30). PPAR␥-mediated apoptosis induction could therefore occur either via stimulation of the proapoptotic effects of TNF␣, or alternatively by interfering negatively with the protective NFB signaling pathway. Interestingly, recently it was suggested that, in murine macrophages, PPAR␥ may inhibit the inflammatory response in part by antagonizing the activities of transcription factors of the AP-1, STAT, and NFB families (38). Therefore it is likely that PPAR␥ induces cellular apoptotic death by interfering negatively with the antiapoptotic NFB signaling pathway. Our results from co-transfection experiments show that PPAR␥ inhibits the transcriptional activity of p65/RelA, a key component of the anti-apoptotic NFB signaling pathway (28). Thus, in addition to the previous described inhibition of transcriptional activity of members of the E2F/DP family (50), interfering with the NFB signaling function constitutes a novel mechanism through which PPAR␥ may interfere with cell cycle control and survival.
The observation that PPAR␣ activators only induce apoptosis of activated macrophages is intriguing and may be of importance for the treatment of atherosclerosis. Although the reasons for this selective effect of PPAR␣, as opposed to PPAR␥ activators are presently unclear, this is unlikely to be due to a low level of PPAR␣ expression, since its expression is higher than PPAR␥ in macrophages in all steps of the differentiation process. Furthermore, recent work from our laboratory in human smooth muscle cells indicates that PPAR␣ can also interfere negatively with the p65/RelA subunit of the NFB signaling pathway (51). The major difference between PPAR␥ and PPAR␣ in macrophages is that PPAR␥ protein is predominantly localized in the nucleus of differentiated macrophages, whereas PPAR␣ resides primarily in the cytoplasmic compartment. Therefore the different subcellular localization of PPAR␣ and PPAR␥ in macrophages may be involved in the differential effects of PPAR␣ and PPAR␥ on macrophage apoptosis. TNF␣ and IFN␥ may modify PPAR activity and cytolocalization through phosphorylation leading to a modified activity on p65/ RelA. Indeed, PPARs are phosphoproteins whose activity has been shown to be modulated through phosphorylation induced by factors as insulin and TNF␣ (52,53).
Both unspecific cell necrosis as well as programmed cell death by apoptosis occur in advanced atherosclerotic lesions (54). Macrophages and smooth muscle cells make up the bulk of apoptotic cells in the atherosclerotic plaque (27,54,55) and proapoptotic proteins, such as caspase-3, are expressed in and colocalize with the apoptotic macrophages, T-lymphocytes, and smooth muscle cells in human atheroma (56). It remains to be determined whether induction of macrophage apoptosis by PPAR activators also occurs in vivo in the atherosclerotic plaque and what the (patho)physiological consequences might be.