Pioglitazone increases apolipoprotein A-I production by directly enhancing PPRE-dependent transcription in HepG2 cells.

Pioglitazone, a hypoglycemic agent, has been shown to increase plasma HDL cholesterol, but the mechanism is incompletely understood. We further investigated effects of pioglitazone on transcriptional regulation of apolipoprotein (apo)A-I gene and functional properties of pioglitazone-induced apoA-I-containing particles. Pioglitazone dose-dependently stimulated apoA-I promoter activities in HepG2 cells. A peroxisome proliferator-activated receptor (PPAR)-response element located in site A (-214 to -192 bp, upstream of the transcription start site) of the promoter is required for pioglitazone-induced apoA-I gene transcription. Deletion of site A (-214 to -192 bp), B (-169 to -146 bp), or C (-134 to -119 bp), which clusters a number of cis-acting elements for binding of different transcription factors, reduced the basal apoA-I promoter activities, and no additional pioglitazone-sensitive elements were found within this region. Overexpression or knock-down of liver receptor homolog-1, a newly identified nuclear factor with strong stimulatory effect on apoA-I transcription, did not alter pioglitazone-induced apoA-I transcription. Pioglitazone-induced apoA-I transcription is mainly mediated through PPARalpha but not PPARgamma in hepatocytes. Pioglitazone induced production of HDL enriched in its subfraction containing apoA-I without apoA-II, which inhibited monocyte adhesion to endothelial cells in vitro. In conclusion, pioglitazone increases apoA-I production by directly enhancing PPAR-response element-dependent transcription, resulting in generation of apoA-I-containing HDL particles with increased anti-inflammatory property.


Materials
Tissue culture materials, media, and FBS were obtained from Sigma Chemical Co. Human hepatoblastoma cell line (HepG2 cells) and human monocytic THP-1 cell line were obtained from American Type Culture Collection. Human aortic endothelial cells (HAEC) were purchased from Lonza Biologics. The polyclonal antibodies for human apoAI and apoAII were purchased from Calbiochem. All other chemicals used were of analytical grade.

Cell culture
HepG2 cells were maintained in DMEM containing 10% FBS supplemented with 100 units/ml penicillin G and 100 g/ml streptomycin sulfate at 37°C in a humidifi ed atmosphere of 95% air, 5% CO 2 . HepG2 cells were incubated in the presence or absence of pioglitazone as indicated. Human monocyte THP-1 cells and HAEC were grown in RPMI1640 (Invitrogen) and EMB2 (Cambrex Bio ScienceWalkersville, Inc.) medium, respectively.

Immunoprecipitation and Western blotting
HepG2 cells were grown in 6-well plates until they attained 75-80% confl uence. HepG2 cells were incubated with various amount of pioglitazone (0-25 M) at 37°C for 24-48 h. At the termination of the incubation, culture medium was collected. A 1 ml sample of culture medium was used to measure apoAI secreted into the media by immunopreciptation and Western blotting using anti-human apoA-I or apoA-II antibodies.

RT and quantitative real-time PCR
Total RNA extracted from cells was reverse transcribed with 100 units of SuperScript II reverse transcriptase (Invitrogen). elevation) is greater than that observed with statins and several fi brates and just below that of niacin ( 3 ).
Initially, glitazones were designed as agonistic ligands of the nuclear receptor peroxisome proliferator-activated receptor (PPAR)-␥ to enhance insulin action mainly in skeletal muscle and adipose tissues to promote glucose utilization and suppress intracellular lipolysis. PPARs (PPAR ␣ , ␥ , and ␦ ) are members of the nuclear receptor family, and by ligand activating and heterodimerization of PPAR with retinoid X receptor and binding to the DNA sequence in the promoter region, modulate transcription of various target genes involved in glucose and lipid metabolism ( 14 ). The effects of pioglitazone to reduce triglycerides are mainly shown to be mediated by increased lipoprotein lipase expression through the activation of PPAR ␥ and decreased plasma apoC-III, an inhibitor for lipoprotein lipase-mediated lipolysis ( 15 ). However, mechanisms of the effect of pioglitazone to raise HDL and apoA-I (the major protein of HDL) are incompletely understood. In initial studies, we have shown that pioglitazone did not change uptake of HDL protein and cholesterol ester but increased apoA-I and A-II de novo synthesis, secretion, and mRNA in human hepatoblastoma cells (HepG2) ( 16 ). These studies suggest that pioglitazone, by increasing hepatic apoA-I mRNA expression, increases apoA-I-containing HDL particles without infl uencing the HDL catabolic pathway in hepatocytes.
ApoA-I is mainly synthesized in the liver and intestine, and its expression levels are highly regulated by a number of transcription factors in response to nutrients, hormones, growth factors, cytokines, and drugs such as fi brates. These trans -acting factors bind to the specifi c DNA sequence elements in the promoter region to negatively or positively modulate the transcription of apoA-I gene ( 17,18 ). Functional analysis of apoA-I gene promoter by mutagenesis identifi ed that roughly three sections, A, B, and C, clustering cis -acting elements located between nucleotides Ϫ 222 and Ϫ 110 upstream from the transcription start site, are essential and suffi cient for liver-specifi c expression of apoA-I gene in hepatocytes [reviewed in ( 19 )]. These cis -acting elements in site A ( Ϫ 214 to Ϫ 192) bind thyroid receptor, hepatic nuclear factor (HNF)-4, PPAR ␣ , RAR ␣ / ␤ , retinoid X receptor-␣ , C/EBP, ARP-1, Rev-erb ␣ , tumor necrosis factor-␣ , and interleukin-1; in site B ( Ϫ 169~ Ϫ 146) bind glucocorticoids, estradiol, and HNF3b; and in site C ( Ϫ 134~ Ϫ 119) bind HNF-4 and ARP-1. More recently, a new nuclear factor, liver receptor homolog-1 (LRH-1), with a strong stimulatory effect on apoA-I gene transcription, has been identifi ed to be located in site C ( 20 ). Site C also has a direct repeat-1 sequence for binding a liver X receptor agonist TO901317 ( 21 ). Moreover, the regulation of the hepatic expression of apoA-I gene is controlled by synergistic interaction between these transcription factors ( 18 ). The present study was designed to further investigate the effects of pioglitazone on transcriptional regulation of apoA-I gene and properties of pioglitazone-induced apoA-I-containing HDL particles. purifi ed by phenol/chloroform extraction after digested with 20 g/ml proteinase K at 55°C for 1 h. DNA samples were resuspended in 30 l TE buffer. The 2 to 4 l samples were used for PCR, which was performed at initial denature at 94°C for 2 min followed by 35 cycles at 94°C for 30s, 60°C for 45s, and 72°C for 1min, followed by 72°C for 7 min, with primers that amplifi ed 264 bp of sequences containing the PPAR-response element (PPRE) site in the human apoA-1 promoter from Ϫ 252 to +12 bp, 5 ′ -CCGGGAGACCTGCAAGCCTGC and 5 ′ -GCACCTCCT-TCTCGCAGTCTC. PCR products were resolved on a 1.2% agarose gel.

Adhesion assay
THP-1 cells were labeled with fl uorescence BCECF (Invitrogen) at 37°C for 30 min. HAEC (0.1-1 × 10 5 cells) were plated in 24-or 48-well plates and preincubated at 37°C for 24 h in the presence or absence of fi lter-concentrated medium ( ‫ف‬ 1.2 mg/ ml protein) from cultured HepG2 cells that had been treated with pioglitazone or vehicle DMSO. The labeled THP-1 cells (2 × 10 5 ) and HAEC were co-incubated at 37°C for 10 min. Nonattached THP-1 cells were removed by gently washing three times with PBS containing 1% FBS. The cells were dissolved in PBS-0.1% SDS and the fl uorescence was measured in NOVOstar (BMG LABTECH GmbH, Offenburg, Germany) with 485 nm excitation and 562 nm emission. Separate wells containing serial diluted labeled THP-1 cells were used for generating a standard curve. The ratio of fl uorescence intensity of the attached cells to that of the total labeled cells applied to the well was calculated, and the adhesion activity of THP-1 monocytes to HAEC is expressed as percent of the number of attached monocytes of the total.
For immunodepletion of apoA-I, the conditioned medium was incubated overnight at 4°C in the presence of rabbit anti-apoA-I antibody (Calbiochem) at dilution of 1:20 or rabbit IgG control. The immunoprecipitated apoA-1 were absorbed on protein A-agarose beads. The supernatants were used in the adhesion assay.

Statistical analysis
Data presented are mean ± SD of three experiments. Student's t -test was utilized comparing to control or between two groups. One-way ANOVA analysis was used for the multiple and full set comparisons where appropriate. A P -value <0.05 was considered signifi cant.

Pioglitazone stimulated apoA-I gene transcription
Pioglitazone (1-25 M) increased apoA-I mRNA ( Fig. 1A ) and protein expression ( Fig. 1B ) in cultured human HepG2 hepatocytes. In transient transfection experiments with the reporter plasmids containing artifi cial PPAR elements, pioglitazone was shown to be a weak PPAR ␣ activator via a direct binding of pioglitazone to PPAR ␣ in nonhepatic cell system ( 22 ). To further test whether pioglitazone directly stimulates apoA-I transcription in hepatocytes, we constructed a luciferase reporter plasmid for the apoA-I promoter comprising the nucleotides Ϫ 470 to +8 upstream of the apoA-I gene transcriptional stat site (+1) ,which is necessary and suffi cient for expression of the apoAI gene in hepatocytes. Results of the luciferase assay showed that pioglitazone dose-dependently (1-25 M) increased the apoA-I promoter activity in HepG2 cells ( Fig. 1C ).

Transfection and dual luciferase assay
HepG2 cells were grown in 24-well or 48-well plates for 18-24 h. Cells were then transfected for 24 h with the pGL3-apoAI-Luc plasmid or pGL3 basic luciferase reporter vector as control, using 1.5 l of FuGENE 6 transfection reagent (Roche). All cells were also cotransfected with 2 ng of phRL-TK (Int-) vector (Promega) containing wild-type Renilla luciferase (R luc ) as an internal control for transfection effi ciency in each well. Twenty-four hour later, the transfected cells were washed with PBS two times, and then cultured for 48 h in DMEM containing 1 mg/ml BSA free of fatty acids and various amounts of pioglitazone as indicated. After washing with PBS, cells were lysed for luciferase assay using the Dual-Luciferase Reporter Assay System according to the manufacturer's protocol (Promega). Luminescence was measured on NOVOstar (BMG LABTECH GmbH, Offenburg, Germany). ApoA-I promoter activities are refl ected by luciferase activities that are expressed as relative light units (RLU) to internal control R luc . Data are expressed as mean ± SD of three experiments.
For small interfering RNA (siRNA) knock-down assay, 0.2 g of pGL3-apoAI-Luc was cotransfected with 50 nM of siRNA (Ambion) in 24-well plates for 24 h and then treated with pioglitazone for 48 h. Luciferase assay was performed as described above.

Chromatin-immunoprecipitation assay
HepG2 cells (5 × 10 6 ), grown in 10 cm dishes or 75 cm 2 fl asks, were detached by trypsin-EDTA and resuspended in 9 ml DMEM medium. Cross-linking of proteins to DNA was performed by adding 270 l of 37% formaldehyde (fi nal 1%), and tubes were rocked gently at room temperature for 10 min. Cross-linking was quenched by adding 1 ml of 1.25 M glycine and gently rocking for 5 min at room temperature. Cells were washed three times with 10 ml ice-cold PBS. Cells were pelleted by centrifugation at 280 g for 5 min at 4°C and resuspended in 400 l chromatinimmunoprecipitation (ChIP) buffer (150 mM NaCl, 50 mM Tris-HCl, pH 7.5, 5 mM EDTA, 0.5% NP40, 1% Triton X-100) with protease inhibitors (complete tablet, Roche, Indianapolis, IN), and lysed on ice for 10 min. Nuclei were collected by centrifugation at 600 g for 5 min at 4°C and dissolved in 400 l of ChIP buffer. Nuclei were resuspended 10 times by using a syringe with #26 needles before sonicating at an output of 40% for 5 s for a total 10 times in a Branson Sonifi er 250 (Newtown, CT). Debris was cleared by centrifugation at 12,000 g for 15 min at 4°C, and the supernatant was split into 100 l aliquots in 1.5 ml microcentrifuge tubes for immunoprecipitation (20 l was saved for preparation of input DNA). Two micrograms of specifi c polyclonal antibodies against the human PPAR ␣ and PPAR ␥ (SC-9000× and SC-7196×, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) or normal rabbit IgG was added to each tube, and tubes were rotated overnight at 4°C. Fifteen microliters of protein A agarose beads (Roche, Inc., Indianapolis, IN) was added to 1.5 ml microcentrifuge tubes and rotated for 1 h at 4°C. Beads were collected by centrifugation at 2,000 g for 30 s at 4°C and then washed fi ve times by removing the supernatant. Immunoprecipitates were eluted in 200 l of 50 mM Tris-Cl (pH8.0), 1 mM EDTA, 1% SDS, and 50 mM NaHCO 3 at 65°C for 10 min two times. Then 21 l of 4 M NaCl was added into 400 l sample (fi nal 200 mM) to decrosslink the protein-DNA complex at 65°C for 5 h. DNA was apoA-I promoter activity when the PPRE binding site was mutated ( Fig. 2C ).
To examine whether there are additional potential pioglitazone-sensitive elements in the apoA-I promoter, we performed a deletion mapping assay. Deletion of site A ( Ϫ 214 to Ϫ 192), site B ( Ϫ 169 to Ϫ 146), and site C ( Ϫ 134 to Ϫ 119) reduced the apoA-I promoter activity by approximately 3-, 6-, and 9-fold, respectively, compared with the full-length promoter, indicating that elements in these regions are important in regulation of its basal transcription activities ( Fig. 3 ). When A, B, and C in each site was deleted individually, the basal promoter activities were also reduced ( Fig. 4 ). No signifi cant stimulatory effect by pioglitazone was found in these experiments. These data indicate that the integrity of all of A, B, and C sites in the apoA-I promoter is crucial for its full transcription activities and no other cis -elements responding to pioglitazone were found in these regions examined.

LRH-1 is not involved in pioglitazone-induced apoA-I transcription
To further test our deletion mapping observation, we chose the LRH-1, a newly identifi ed transcription factor located in site C ( Ϫ 133 to Ϫ 117), which strongly stimu-PPRE is required for pioglitazone-induced apoA-I promoter activity PPAR ␣ , but not PPAR ␥ , is highly expressed in the liver. Agonists for PPAR ␣ , such as fenofi brate, have been shown to increase apoA-I gene expression ( 14 ). Although pioglitazone has been previously shown to weakly activate PPAR ␣ , the potential functional consequence has not been tested in hepatocytes. A PPAR ␣ binding sequence, the PPRE, is present in site A of the apoA-I promoter (TGAACCCTTGACCCC). Cotransfection with either PPAR ␣ or PPAR ␥ increased the apoA-I promoter activities, suggesting that activation of both PPAR ␣ and PPAR ␥ are able to stimulate apoAI transcription ( Fig. 2A ). Mutating the PPRE sequence by changing two nucleotides, A to T, dramatically reduced the promoter activity, indicating the importance of this binding element in the regulation of apoA-I expression in hepatocytes ( Fig. 2 ). Transfecting HepG2 cells with PPAR ␣ and PPAR ␥ did not reverse the mutated apoA-I promoter activity ( Fig. 2A ). Furthermore, these mutations abolished pioglitazone-induced apoA-I promoter activity, suggesting that the PPRE is required for the effect exerted by pioglitazone ( Fig. 2B ). Similarly, a PPAR ␣ ligand and activator, fenofi brate, signifi cantly stimulates apoA-I transcription, but it did not reverse the glitazone, suggesting that both PPAR ␥ ligand and PPAR ␣ agonist are able to bind the same PPRE in the apoA-I promoter (luciferase activity fold change: 1.39 ± 0.1 for 10 M pioglitazone, 1.81 ± 0.39 for 10 M WY14643, and 1.37 ± 0.22 for 10 M pioglitazone plus 10 M WY14643). We then performed siRNA knock-down assays. siRNA knocking-down reduced mRNA expression of PPAR ␣ by 44% (0.56 ± 0.087 vs. 1.0 ± 0; P < 0.01) and PPAR ␥ by 40% (0.60 ± 0.048 vs. 1.0 ± 0; P < 0.01), respectively (RT-qPCR fold change). As a positive control for our silencing assays, knocking-down PPAR ␣ reduced fenofi brate-induced apoA-1 transcription by 74% (lucifease activity: 39.5 ± 3.1 vs. 14.4 ± 0.41; P < 0.01). Knock-down of PPAR ␣ eliminated pioglitazone-induced apoA-I transcription ( Fig. 6A ), whereas knock-down of PPAR ␥ largely retained pioglitazone-induced apoA-I transcription ( Fig. 6B ), suggesting that PPAR ␣ but not PPAR ␥ plays a major role in pioglitazone-induced apoA-I expression in hepatocytes.
To further examine the role of PPAR ␣ and PPAR ␥ in pioglitazone-induced apoA-I transcription, we performed a ChIP assay. The apoA-I promoter sequence from Ϫ 252 to +12 bp containing the PPRE element was amplifi ed by PCR in the DNA/chromatin fragments immunoisolated with anti-PPAR ␣ antibody but not with anti-PPAR ␥ antibody ( Fig. 7 ). The same apoA-I promoter region was amplifi ed by PCR from the samples treated with Wy14643, a lates apoA-1 expression in the liver ( 20 ). Overexpression of LRH-1 increased apoA-I promoter activity; addition of pioglitazone did not further increase the promoter activity ( Fig. 5A ). Furthermore, siRNA knocking-down LRH-1 reduced LRH-1 mRNA expression by 55% [RT and quantitative real-time PCR (RT-qPCR) fold change: 0.443 ± 0.145 vs. control 1.0 ± 0; P < 0.01) and decreased apoA-I transcription, but did not affect pioglitazone-induced activity, as pioglitazone similarly increased apoA-I transcription by 162% and 168% in control and siRNA-treated cells, respectively, suggesting that LRH-1, with its binding element in site C, is not involved in pioglitazone-induced apoA-I transcription ( Fig. 5B ).
To investigate the role of PPAR ␣ and PPAR ␥ in pioglitazone-induced apoA-I transcription, we fi rst tested the effects of a PPAR ␣ agonist together with pioglitazone. No additive effects were observed when cells were treated with pioglitazone alone and PPAR ␣ agonist YW14643 plus pio- the effect of apoA-I-containing HDL particles secreted by HepG2 cells on monocyte adhesion to HAEC, a critical initial infl ammatory event involved in monocyte accumulation. As shown in Fig. 9A , the conditioned media (CM) from pioglitazone-treated (10 M pioglitazone) HepG2 cells showed a stronger inhibitory effect on monocyte adhesion than the CM without pioglitazone treatment (0 M pioglitazone, as treated with vehicle DMSO) ( P = 0.038), suggesting an enhanced anti-infl ammatory activity. During the adhesion assays, incubation of these CM for 24 h did not alter HAEC viability and/or proliferation as measured by CellTiter-Glo Luminescent Cell Viability kit (Promega; this assay is a highly sensitive method for assessing cell proliferation and cytotoxicity). The RLU for the control CM and pioglitazone-treated CM were 56,140 ± 482 and 60,717 ± 1,772, respectively.
To further test whether apoA-I in the CM is directly responsible for the reduced adhesion activity, we performed immunodepletion of apoA-I and compared the effects of CM containing apoA-I (CM apoA-1+) and CM without apoA-I (CM apoA-1-). As shown in Fig. 9 , the CM apoA-1 Ϫ , from either the 0 M pioglitazone-treated or 10 M pioglitazone-treated HepG2 cells, reduced their ability to inhibit the adhesion ( P = 0.006 and P = 0.016). Comparisons between groups in CM apoA-1+ and control were significant ( P for trend = 0.001) but not signifi cant between groups in CM apoA-1and control. These results suggest that apoA-1/apoA-1-containing particles signifi cantly con-PPAR ␣ agonist that served as a positive control. These results clearly show that pioglitazone, by mainly interacting with PPAR ␣ but not PPAR ␥ , increased apoA-I gene transcription in hepatocytes.

Pioglitazone-induced apoA-I production is mainly in LP-A-I particles
In HepG2 cells, we have previously shown that mainly two HDL particles are secreted, including HDL particles containing apoA-I only (LP-AI) or HDL particles containing both apoA-I and apoAII (LP-AI+AII) ( 16 ). To assess the distribution of pioglitazone-induced apo-A-I in these particles, we analyzed the secreted apoAI-containing particles by immunoprecipitation and Western blotting. As shown in Fig. 8 , pioglitazone (1-10 M) mainly increased apoA-I in LP-AI particles by 1.2-to 1.4-fold, and, to lesser degree, in LP-AI+AII particles by 1.1-to 1.2-fold ( Fig. 8 ).
These data indicate that the increased apoA-I by pioglitazone was mainly enriched in LP-AI particles.

Conditioned media from pioglitazone-treated HepG2 cells increased the inhibition of monocyte adhesion to HAEC
In previous studies, we have shown that pioglitazoneinduced apoA-Icontaining particles were able to mediate cholesterol effl ux from macrophages ( 16 ). To assess antiinfl ammatory biological properties of HDL particles generated from pioglitazone-treated HepG2 cells, we tested Among the three synthetic glitazones with PPAR-agonist activities, pioglitazone has a much better effect on raising HDL and lowering triglycerides. The mechanism for this divergence has been suggested to be due to the ability of pioglitazone to weakly activate PPAR ␣ . In transient transfection experiments with the reporters containing four copies of the rat acyl-CoA oxidase PPRE, previous studies showed that pioglitazone appeared to be a weak PPAR ␣ activator via a direct binding of pioglitazone to the artificial PPRE in kidney-derived COS-1 cells ( 22 ). There is one PPRE in site A of the apoA-I promoter, and the PPAR ␣ expression level is much higher than PPAR ␥ in the liver ( 14 ). We therefore hypothesized that a weak binding activity of pioglitazone to PPAR ␣ observed in a nonhepatocyte system would have the potential to initiate transcription of apoA-I gene in hepatocytes, resulting in an increased production of apoAI-containing HDL particles in the liver. In this study, we further assessed the effects of pioglitazone on apoAI transcriptional regulation in HepG2 cells. Using transfection studies in HepG2 cells with full-length apoA-I promoter, we showed that pioglitazone stimulated apoA-I transcription in HepG2 cells, a homogeneous background for its expression.
Furthermore, our studies using deletion mapping and site-directed mutagenesis have defi ned that a single PPAR ␣ tribute to the inhibitory effects on the monocyte adhesion to the endothelial cells.

DISCUSSION
A number of clinical studies established that pioglitazone signifi cantly increased serum HDL-cholesterol and decreased triglycerides, and pioglitazone produced more favorable lipid profi les than rosiglitazone in patients with type 2 diabetes (23)(24)(25)(26). However, the effect of pioglitazone on the plasma apoA-I level has not been studied in detail. In a small number of nondiabetic patients with low HDL-cholesterol and metabolic syndrome, Szapary et al. ( 24 ) have recently shown that pioglitazone treatment for 6 weeks signifi cantly increased mean apoA-I by 6.8%. Recently, Davidson et al. ( 26 ) also reported that pioglitazone treatment for 24 weeks increased serum apoA-I in patients with type 2 diabetes. In an attempt to understand the mechanisms of apoA-I increase by pioglitazone, using human hepatoblasoma cell line (HepG2 cells), we have previously shown that pioglitazone increased the de novo synthesis and secretion of apoA-I particles in the culture medium by increasing apoA-I mRNA levels without affecting HDL catabolic events in HepG2 cells ( 16 ). with individually deleted regulatory sites A, B, and C in the promoter. In these studies, we were not able to identify the possibility of any other cis -elements in pioglitazone-mediated ApoA-I transcription in hepatocytes. Although our data suggest that PPRE is importantly linked to apoA-I transcription by pioglitazone, this study cannot completely exclude other factors that may participate in pioglitazoneinduced apoA-I transcription. Because site B is mainly involved in hormonal regulated apoA-I transcription, it is unlikely that site B plays a major role in pioglitazonemediated apoA-I transcription. Overexpression or knocking-down of a strong apoA-I activator LRH-1, a binding sequence located in site C ( 20 ), did not alter pioglitazoneinduced apoA-I promoter activity. Furthermore, chromatin immunoprecipitation showed that pioglitazone increased the binding of PPAR ␣ to the apoA-1 promoter containing PPRE in vivo ( Fig. 7 ). Our data indicate that the PPRE could be a major functional site for pioglitazone-induced apoA-I transcription in hepatocytes.
Previous studies have shown that pioglitazone is a PPAR ␥ agonist but also a weaker PPAR ␣ activator; both PPAR ␥ binding element located in site A of the promoter region is required for pioglitazone-induced apoA-I expression in the hepatocyte system. In hepatocytes, we suggest that pioglitazone exerts its activating effects on apoA-I gene through a PPRE-dependent event. A modest stimulatory effect on apoA-I transcription reported in this study is consistent with a weaker binding activity of pioglitazone to PPAR ␣ observed in nonhepatic cells. In additional studies, we investigated the potential role of pioglitazone in the synergistic integrity of all three clustering cis -acting elements of A-, B-, and C-sites within the apoA-I gene in regulating apoA-I transcription in hepatocytes. Deletion of all sites A, B, or C individually markedly reduced basal apoA-I promoter activity, even when the PPRE in site A remained in the promoter. Our data suggest that the integrity of all three cis -acting element sites A, B, and C are essential and required for the full apoA-I transcription in HepG2 cells, similar to the previous reports ( 21 ). To further identify the possibility of additional cis -elements that may be operating in pioglitazone-induced apoA-I transcription, we assessed the effect of pioglitazone on apoA-I transcription apoA-I and plasma apoA-I levels as a compensatory mechanism through PPAR ␥ when PPAR ␣ in the liver is suppressed or PPAR ␥ expression is enhanced in the liver, such as in several murine models of obesity or diabetes ( 27 ).
To assess properties of HDL particles secreted from HepG2 cells that were treated with pioglitazone, we further analyzed apoAI-containing particles. We found that pioglitazone-induced apoA-I production is mainly enriched in LP-AI particles, and these apoAI-containing particles are more effective in inhibiting adhesion of THP-1 monocytes to HAEC ( Figs. 8 and 9). These fi ndings indi-and PPAR ␣ can bind the PPRE in the promoter region of human apoA-I gene and pioglitazone can effectively compete for these bindings ( 22 ). Data from our cotransfection and siRNA knock-down studies and from other reports ( 16,19,22 ) indicate that pioglitazone, through activation of PPAR ␣ , increases apoA-I transcription and protein production, as PPAR ␣ is predominately expressed. Our ChIP assay data showed that pioglitazone-induced apoA-1 expression is mainly through the direct binding of PPAR ␣ but not PPAR ␥ to PPRE element in apoA-I promoter in hepatocytes. In vivo studies would be helpful to further test whether pioglitazone will increase expression of  HepG2 cells were grown to 80% confl uence and then incubated in the presence of either vehicle (DMSO) or 10 uM of pioglitazone and Wy14643 for 48 h. ChIP were performed as described in "Materials and Methods." Input: sheared nuclear lysates before subjecting to immuoprecipitation using: No Ab, rabbit IgG control, PPAR ␣ , anti-PPAR ␣ antibody, PPAR ␥ , anti-PPAR ␥ antibody. Fig. 8. Pioglitazone increased apoA-1 in LP-AI particles from HepG2 cells. HepG2 cells were incubated with pioglitazone for 48 h. LP-AI particles containing only apoA-1 (apoA-1 only HDL) were isolated from medium by immunoprecipitation fi rst with apoAII antibody to remove apoAII-containing particles and then with apoAI antibody. LP-AI+AII particles (Total HDL) were isolated by immunoprecipitation with both apoAI and apoAII antibodies. Immunoprecipitates were then examined by Western blotting with apoA-1 antibody. Representative Western blots are shown with fold changes compared with control by densitometrical quantitation of three blots. lism ( 28,29 ). This information is important in combination therapy using drugs with complementary mechanisms of action to achieve aggressive HDL goals. Additional research is needed to assess the additive effi cacy of such combinations not only on lipids, but also on reducing cardiovascular events in the high risk population of diabetics beyond monotherapy.
In summary, we suggest that pioglitazone, through PPRE-mediated events within cis -acting element site A in the apoA-I gene promoter, increases hepatic apoA-I mRNA expression, resulting in increased plasma levels of physiologically active apoA-I/HDL particles. We also suggest that the enhanced reverse cholesterol transport and antiinfl ammatory properties of apoAI-enriched HDL particles generated by pioglitazone in hepatocytes may, at least in part, explain the cardio-protective actions of pioglitazone in inhibiting atherosclerosis in humans ( 26 ). It is also important to note that the effective concentrations of pioglitazone used in our studies (0.1-25 mol/L) are comparable cate that pioglitazone, by enhancing PPRE-dependent apoAI transcription, increased the de novo synthesis and secretion of apoAI-enriched LP-AI particles in the culture medium, which in turn participated in signifi cantly inhibiting the adhesion of monocytes to aortical endothelial cells, a key initial infl ammatory event associated with monocyte migration and accumulation during atherogenesis. Based on our previous and this study, we suggest that the pioglitazone-induced production of functionally active apoAI-containing HDL particles, by enhancing reverse cholesterol transport pathway and anti-infl ammatory properties (two important benefi cial functions of HDL), inhibit atherogenesis.
Although originally developed as an agent to control glycemia in diabetic patients, pioglitazone should also be considered as part of dyslipidemia treatment in diabetes in combination with lipid-regulating drugs including statins, niacin, and fi brates. Whereas fi brates and statins act via PPAR-activation, niacin mainly decreases apoA-I catabo- Fig. 9. Conditioned medium from pioglitazone-treated HepG2 cells inhibited monocyte adhesion to HAEC. A: HepG2 cells were incubated with DMSO vehicle (0 M) or 10 M pioglitazone for 48 h. The CM was fi lter concentrated and washed with PBS in an Amicon ultra centrifugal tube, Ultracel-50k (Amicon). The concentrated CM was then incubated 16-24 h at 4°C in the presence of rabbit anti-apoA-1 antibody (Calbiochem) at a dilution of 1:20 or in the presence of rabbit IgG as control for immunodepletion of apoA-1. The immunoprecipitated apoA-1 were removed by the binding to protein A-agarose beads. The supernatants were used in the adhesion assay of the fl uorescence-labeled THP-1 monocytes to endothelial cells, as described in "Materials and Methods." Control: The coincubation of labeled THP-1 cells and HAEC in assay (basal EMB2) medium; CM apoA-1+: the CM after immunoprecipitation with rabbit IgG control; CM apoA-1-: the conditioned medium after immunodepletion of apoA-1 with anti-apoAI. Then 1 g purifi ed apoAI was added as positive control in the adhesion assay. Data are expressed as mean ± SD of three experiments. One-way ANOVA analysis was used for the multiple comparisons between groups and control. B: Western blotting analysis of apoA-1 in the conditioned medium: CM apoA-1+ and CM apoA-1-.