Hepatocyte Growth Factor Regulates E Box–Dependent Plasminogen Activator Inhibitor Type 1 Gene Expression in HepG2 Liver Cells
Objective— We sought to determine the etiologic mechanism of pleiotropic growth factor, hepatocyte growth factor (HGF), as a regulator of hepatic synthesis of plasminogen activator inhibitor (PAI)-1, the physiological inhibitor of fibrinolysis and a potential inducer of atherothrombosis.
Methods and Results— HGF increased PAI-1 mRNA expression and PAI-1 protein accumulation in the conditioned media of human liver-derived HepG2 cells, and increased hepatic PAI-1 mRNA expression in vivo in mice. HGF-inducible PAI-1 mRNA was attenuated by U0126, a specific inhibitor of mitogen-activated protein kinase (MAPK) kinase, and genistein, an inhibitor of tyrosine kinase. HGF increased the human PAI-1 promoter (−829 to +36 bp) activity, and deletion and mutation analysis uncovered a functional E box (5′-CACATG-3′) at positions −158 to −153 bp. Electrophoretic mobility shift assays demonstrated that this E box binds upstream stimulatory factors (USFs). HGF phosphorylated USFs through MAPK and tyrosine kinase pathways. Co-transfection of USF1 expression vector increased PAI-1 promoter activity. Sterol regulatory element-binding protein-1 attenuated HGF-inducible PAI-1 promoter activity.
Conclusions— Because USFs are involved in the regulation of carbohydrates and lipid metabolism, HGF-mediated PAI-1 production may provide a novel link between atherothrombosis and metabolic derangements. Targeting HGF signaling pathway may modulate the thrombotic risk in high-risk patients.
Hypofibrinolysis mediated by increased circulating PAI type 1 (PAI-1), the major physiological inhibitor of fibrinolysis, confers thrombotic risk.1 Hemostatic and fibrinolytic system abnormalities are also associated with increased cardiovascular morbidity and mortality in insulin resistance.2
Hepatocyte growth factor (HGF) is an endothelial growth factor with potent angiogenic and mitogenic properties. Increased concentrations of HGF in blood portend progression of cardiovascular diseases.3 Patients with acute coronary syndromes and atherosclerosis exhibit increased HGF concentrations in blood.4,5 HGF is increased also in insulin resistance.6 3T3-L1 adipocytes can secrete HGF in vitro,7 and adipose tissue may be a source of HGF in blood. HGF increases PAI-1 expression in hepatocytes in vitro.8 Thus, HGF may influence fibrinolysis and atherosclerosis by altering PAI-1 expression.
Mechanisms responsible for increased PAI-1 with insulin resistance are not fully elucidated.9 Adipose tissue may be an important source of PAI-1 with excess fat and subclinical inflammation.10 However, fatty liver is common in metabolic syndrome,11 and liver steatosis is strongly related to plasma PAI-1 levels,12 suggesting that liver is an important source of increased circulating PAI-1. This study was designed to characterize molecular mechanisms influencing altered hepatic PAI-1 expression by HGF and to identify potential therapeutic approaches for attenuation of HGF-induced PAI-1 expression.
HepG2 cells, a highly differentiated human hepatoma cell line (American Type Culture Collection, Rockville, Md), were grown in Dulbecco’s modified Eagle medium (DMEM; Sigma, St. Louis, Mo) supplemented with 10% fetal bovine serum (Sigma).13 At 80% confluence cells were washed with phosphate-buffered saline (PBS) and incubated in serum-free DMEM for 16 hours. The medium was replaced with fresh serum-free DMEM containing recombinant human HGF (R&D Systems, Minneapolis, Minn). When U0126 (Promega, Madison, Wis), genistein (Sigma), GF109203X (Tocris, Ellisville, Mo), and LY294002 (Biomol, Plymouth Meeting, Pa) were used, cells were pretreated with these agents for 15 minutes before HGF exposure. Total protein in the media was determined by the Bradford method. Nuclear extracts were prepared with N-XTRACT kit (Sigma). Where noted, nuclear extracts were incubated with calf intestinal alkaline phosphatase (CIP; New England BioLabs, Ipswich, Mass).
Animals and Treatment
Experiments conformed to the guidelines of Hokkaido University and were approved by Institutional Animal Study committee. Male C57BL/6J mice (CLEA Japan, Tokyo) had free access to chow and water. At 8 weeks (20 to 22 g) mice were given HGF or saline. After 4 hours anesthesia was induced with ether. Livers were excised, rinsed with PBS and snap-frozen in liquid nitrogen for RNA extraction.
Western Blotting and Immunoprecipitation
Equal amounts of proteins were separated by SDS-polyacrylamide gel electrophoresis (PAGE) under reducing conditions on an 8% (acrylamide:bis-acrylamide=29:1 for conditioned media) or 12% (acrylamide:bis-acrylamide=200:1 for nuclear extracts) gel. Western blotting was performed as previously described.13 Immunologic detection was performed with ProtoBlot system (Promega). The primary antibodies were mouse anti-human PAI-1 monoclonal antibody (American Diagnostica, Greenwich, Conn), mouse anti-human vitronectin (V-7881; Sigma), rabbit anti-upstream stimulatory factor (USF)1 (sc-8983; Santa Cruz Biotechnology, Santa Cruz, Calif), and rabbit anti-USF2 (sc-862; Santa Cruz). Secondary antibodies were AP-conjugated rabbit anti-mouse IgG (Sigma) and goat anti-rabbit IgG (sc-2007; Santa Cruz). Blots were analyzed with the use of a Scion Image System (Scion Corporation, Frederick, Md). Immunoprecipitation of nuclear extracts was performed as previously described14 using specific anti-UFS1 antibody. Reactions were resolved by SDS-PAGE and transferred to polyvinylidene fluoride membranes. Membranes were incubated with anti-phosphoserine antibody (Sigma) and immunoreactive proteins were visualized by incubating with horseradish peroxidase-linked sheep anti-mouse secondary antibody.
RNA Isolation, Real-Time Polymerase Chain Reactions, and Northern Blotting
Total RNA was extracted from mouse liver and cells by ISOGEN reagent (Nippon Gene, Tokyo, Japan). For real-time PCR RNA (20 ng/10 μL reaction volume) was used for first strand cDNA synthesis with TaqMan Gold RT-PCR Kits (Applied Biosystems, Foster City, Calif). The extent of PAI-1 gene expression was determined with Assay-on-Demand Gene Expression Assay mixes (Mm00435860_m1; Applied Biosystems) that consisted of specific primers and FAM dye-labeled TaqMan MGB probes and the ABI Prism 7000 Sequence Detector system. GAPDH was used as reference gene. The relative concentration of PAI-1 was adjusted for the corresponding glyceraldehyde 3-phosphate dehydrogenase (GAPDH) concentration. Northern blotting was performed as described previously.13 The membranes were hybridized with human PAI-1 and GAPDH cDNA probes labeled with [α-32P]dCTP (Amersham). The membranes were exposed at −80°C for autoradiography.
Promoter-Luciferase Vector and Expression Plasmid
Human PAI-1 promoter 5′ flanking region from −829 to +36 (865 bp)13 was subcloned into multiple cloning sites of pGEM-T Easy vector (Promega). The T-vector was digested with KpnI and XhoI and subcloned into the KpnI/XhoI sites of the promoterless luciferase reporter gene vector pGL3-Basic (Promega) and referred to as pGL3PAI- 829. Basal luciferase activity was detected by pGL3PAI- 829 vector. The progressive deletion mutants from the 5′ end of the human PAI-1 promoter fragments were generated by PCR with the 5′ primers complementary to the PAI-1 gene sequence. All fragments had an identical 3′ end (ie, the XhoI site at portion +36) and 5′ end (ie, the KpnI site at each portion). Each deletion mutant of PAI-1 promoter vectors was referred to as follows: pGL3PAI-663, pGL3PAI-539, pGL3PAI-366, pGL3PAI-308, pGL3PAI-210, pGL3PAI-171, and pGL3PAI-120. Site-directed mutagenesis of the proximal E box motif (5′-CACATG-3′) at positions −158 to −153bp of pGL3PAI- 829 was performed as previously described15 with an oligonucleotide primer: 5′-GCACACACACACACACACTTAATCCTCAGCAAGTCCC-3′. The mutant vector has 4 point mutations in the E box motif as underlined: 5′-CTTAAT-3′, and is referred to as mutpGL3PAI- 829.
DNA Transfection and Luciferase Assays
HepG2 cells (1.0×106 cells) were inoculated on 35-mm dishes and preincubated in DMEM with 1% fetal bovine serum for 24 hours. Transient transfection was performed with the calcium-phosphate precipitation method. Transfections were performed with 2 μg each of the PAI-1 promoter firefly luciferase fusion DNA reporter construct and 0.5 μg phRG-B vector (Promega), a promoter-less Renilla luciferase reporter plasmid, to control for transfection efficiency. These cells were cultured in serum-free DMEM for 24 hours and stimulated with HGF (100 ng/mL) in serum-free DMEM for 24 hours. Luciferase activity was detected in cell extracts by Dual-Luciferase Reporter Assay (Promega). Normalized luciferase activity was calculated as the ratio of firefly luciferase activity to control Renilla luciferase activity. Results for each reporter construct were expressed as percent induction compared with results in transfected unstimulated cells. The expression vector pCX-USF (kindly provided by RG Roeder, Rockfeller University, NY) encoded for human USF1.16 The expression vector pCMVhSREBP-1a, -1c, and -2 (kindly provided by H. Shimano, Tsukuba University, Ibaraki, Japan) encoded for the human sterol regulatory element-binding protein (SREBP)-1a, -1c, and -2, respectively.17 The total amount of DNA was adjusted equally using empty vector DNA.
Electrophoretic Mobility Shift Assays
Equal amounts of complementary synthetic oligonucleotides containing the E box motif (underlined) (5′-ACACACACACATG CCTCAGC- 3′) were annealed and labeled by 5′ end labeling with [γ-32P]ATP and T4 polynucleotide kinase (Takara, Kyoto, Japan). Unlabeled annealed oligonucleotides containing a mutated E box motif flanked by PAI-1-specific sequence (5′-ACACACACTTAATCCTCAGC-3′) and consensus E box motif (underlined) from adenovirus major late promoter (5′-AGGTGTAGGCCACGTGACCGGGTGTT-3′) were used as competitors. Binding reactions were performed with 10 μg of nuclear extracts. For competition analysis 100-fold molar excess of unlabeled oligonucleotide was added. After preincubation for 20 minutes at room temperature, 1 μL of the labeled probe was added and the incubation was continued for an additional 20 minutes. For supershift analysis 2 μL of USF1 (H-86), USF2 (C-20) and hypoxia inducible factor (HIF)1α (H-206) TransCruz antibodies (Santa Cruz Biotechnology) were added to the reactions and incubated for 1 hour on ice before addition of labeled probe. DNA-protein complexes were separated by electrophoresis on 4% nondenaturing polyacrylamide gel in 0.5× Tris-borate-EDTA buffer. Gels were dried and exposed to a film with an intensifying screen.
Data are means±SD. Differences were assessed by analysis of variance with Bonferroni least significant post hoc tests for comparisons within multiple groups. Significance was defined as P<0.05.
Effects of HGF on Expression of PAI-1 in HepG2 Cells and Mice Liver
HGF increased accumulation of PAI-1 protein in the conditioned media in a concentration-dependent fashion (2.3±1.3-fold at 0.5 ng/mL, 3.5±1.3 fold at 5 ng/mL, 7.9±2.7 fold at 50 ng/mL, and 11.1±5.6 fold at 100 ng/mL, at 12 hours, n=3, Western blot, Figure 1A). Intravenous injection of HGF (100 μg/kg) increased hepatic PAI-1 mRNA expression in mice by 1.6-fold over control (Figure 1B). Vitronectin and total protein content in media were not affected (Figure 1A and 1C).
Induction of PAI-1 mRNA Expression by HGF Mediated by Tyrosine Kinase and MEK-Dependent Signaling Pathways
PAI-1 mRNA expression was increased in cells treated with HGF (Figure 2A). Both 3.2-kb and 2.2-kb forms of PAI-1 mRNA were increased (1.7±0.4-fold at 0.5 ng/mL, 3.3±0.9-fold at 50 ng/mL concerning total PAI-1 mRNA, n=3). To delineate the intracellular signal transduction pathways cells were treated for 15 minutes with each of several selected inhibitors before HGF addition (Figure 2B). The increased PAI-1 mRNA induced by HGF was attenuated by pretreatment with U0126, a specific inhibitor of mitogen-activated protein kinase (MAPK) kinase, and genistein, an inhibitor of tyrosine kinase. By contrast, GF109203X, an inhibitor of the protein kinase C pathway, and LY294002, an inhibitor of phosphatidylinositol 3 (PI-3)-kinase, had no significant effects. Concentrations of GAPDH mRNA did not change in any of these experimental conditions.
Determination of DNA Regions Critical for HGF-Inducible PAI-1 Transcriptional Activity
To identify 5′ flanking region responsible for HGF effects, transient transfections with several PAI-1 promoter-luciferase reporter constructs were performed (Figure 3A). Higher basal promoter activity was observed in pGL3PAI- 210, suggesting that the region between −210 and −308 bp may contain a repressor element. HGF increased promoter-driven luciferase activity by 2.0±0.3-fold (n=5). Relative to the largest promoter fragment tested, HGF effect was reduced with deletion of the region at −171 to −120 bp. These data indicate that the major sequence determinant of responsiveness resides between −171 and −120 bp. Through sequence analysis we identified a potential E box element (5′-CACATG-3′) at −158 to −153 bp that is strongly homologous with the canonical E box. To further characterize the region the mutant construct containing a 4-nucleotide substitution in the E box was generated. Compared with the HGF-induced promoter activity in the wild-type, substitution in the E box reduced HGF-induced PAI-1 promoter activity (Figure 3B). These results suggest that this E box is an HGF responsive sequence.
USFs Bind Specifically to E Box in the PAI-1 Promoter
Nuclear extracts from quiescent and HGF-treated cells were tested in an electrophoretic mobility shift assay (EMSA) with the use of 3 probes: a USF consensus E box as the control probe, a wild-type PAI-1 E box-containing probe, and a mutated PAI-1 E box-containing probe. The results showed an increase in nucleoprotein binding with extracts from HGF-treated cells (Figure 4A). Results of competition experiments indicated that the complex bands were competed out exclusively by adding excess unlabeled probe containing the putative E box motif. No competition occurred when excess unlabeled probe with mutations in the putative E box motif was added. The complex bands were also competed out by addition of the probe containing the E box consensus sequence flanked by non–PAI-1 promoter sequences (Figure 4B). Because USFs were identified initially based on their ability to bind to the E box, we sought to determine whether these transcriptional factors are involved in PAI-1 gene regulation. The anti-USF1 or anti-USF2 antibodies disrupted the band and resulted in a supershift (Figure 4B). When nuclear extracts were incubated simultaneously with anti-USF1 and anti-USF2 antibodies, the supershifted band exhibited slower mobility, indicating that USF proteins bind to the E box as heterodimers of USF1 and USF2. However, no supershift was observed with anti-HIF1α antibody. HIF1 is known as a transcription factor that binds to E box and plays a crucial role in hypoxia-induced expression of PAI-1. The results indicate that an intact hexanucleotide E box is necessary and sufficient for binding of USF1 and USF2. In HGF stimulated cell nuclear extracts, an analogous result was obtained (Figure 4C), suggesting that HGF effects were mediated by USFs.
HGF-Induced Phosphorylation of USFs Mediated by the MAPK Signaling Pathway
The phosphorylation of USFs is necessary for DNA binding. To determine whether USFs are targets for HGF-responsive kinase, the nuclear extracts from HGF-treated HepG2 cells were examined by Western blot with specific antibodies to USF1 and USF2 (Figure 5A). The quiescent cells provided predominant band of 43 kDa USF1 and 44 kDa USF2. HGF treatment induced the appearance of other bands with slower mobility. These slower migrating species correspond to the phosphorylated forms of USF1 (44 kDa) and USF2 (45 kDa), respectively. Both USF1 and USF2 were phosphorylated by HGF. CIP treatment of the nuclear extracts from HGF-treated cells before SDS-PAGE significantly decreased the appearance of these species (Figure 5B). Results of immunoprecipitation reveled that the levels of phosphorylated USF1 were low in unstimulated cells but increased after HGF treatment (Figure 5C). To determine whether the phosphorylation of USF1 was mediated by c-Met tyrosine kinase and MEK/extracellular signal regulated kinase (ERK) pathway, cells were pretreated for 15 minutes with U0126 and genistein before HGF treatment. HGF-induced phosphorylation of USF1 was attenuated by U0126 and genistein (Figure 5D). LY294002 had no effects.
Effects of USF and SREBP on HGF-Induced Expression of PAI-1
Because USF recognizes E box sequences, the involvement of USF in PAI-1 regulation was studied. Co-transfection of cells with USF1 expression vector increased PAI-1 promoter activity. Promoter activity was not increased when mutations were added in the putative E box motif (Figure 6A). HGF-inducible promoter activity was attenuated by co-transfection with pCMVhSREBP-1a and -1c, whereas pCMVhSREBP-2 had no effect (Figure 6B). To investigate the interaction between USF1 and SREBP-1a on PAI-1 promoter activity, the inhibitory effects of SREBP-1a was compared with different amounts of USF1. Basal PAI-1 promoter activity was not affected by SREBP-1a. PAI-1 promoter activity induced by USF1 was repressed by SREBP-1a in a dose-dependent manner, suggesting that SREBP-1a can interfere with USF1 in activating PAI-1 promoter (Figure 6C).
Altered PAI-1 expression likely contributes to cardiovascular derangements, and normalization may be important in high-risk populations. We found that HGF enhances PAI-1 expression in human liver-derived HepG2 cells in vitro and in mice in vivo. HGF at a dose simulating clinically encountered concentrations3 increased PAI-1 accumulation in the conditioned media in a concentration-dependent fashion. Total protein concentration was unaltered by HGF consistent with a previous report.8 Concentrations of vitronectin that stabilizes PAI-1 were not altered, suggesting that HGF effect was specific to PAI-1. Furthermore, HGF increased hepatic PAI-1 in mice in vivo, implicating the physiological relevance of the finding in HepG2 cells in vitro. PAI-1 can be produced in hepatocytes,13 endothelial cells18 and adipocytes.10 Hepatic PAI-1 overproduction can cause systemic fibrinolytic system shutdown as shown by consequences of induction of the MET oncogene in mouse liver resulting in increased hepatic PAI-1 production and disseminated intravascular coagulation.19
Binding of HGF to its specific receptor c-Met induces activation of intracellular tyrosine kinase domain of the c-Met and recruits a number of intracellular signaling molecules. HGF-induced pleiotropic responses are mediated through both MAPK and PI-3 kinase pathways.3 In the present study the increase of PAI-1 mRNA by HGF was attenuated by U0126 and genistein, suggesting that the phosphorylation of ERK1/2, the signaling event downstream of c-Met tyrosine kinase activation, is involved in HGF-induced PAI-1 expression. USF proteins require phosphorylation for DNA binding.20 In this study HGF can phosphorylate USFs. The reduction of phosphorylated USF1 by U0126 and not by LY294002 suggests that USF proteins may be the downstream targets of activated ERK1/2. Peak USF phosphorylation in response to HGF was delayed compared with the reported peak ERK1/2 phosphorylation.21
Deletion and mutation analysis of the PAI-1 promoter and EMSA suggested that signaling involves binding of USF1/2 to the proximal E box located at −158 to −153 bp upstream of the transcription start site of the human PAI-1 gene. USFs are ubiquitous members of the basic-helix-loop-helix family of transcriptional factors. First identified for their involvement in transcription from the adenovirus major late promoter, USF proteins were purified as 2 polypeptides of 43 and 44 kDa, termed USF1 and USF2, respectively. USF proteins bind to the hexanucleotide E box motif (5′-CANNTG-3′) as homo-/hetero-dimers of USF1 and USF2.22 Luciferase assay of the PAI-1 promoter suggested that the region from - 663 to - 539 bp and the region from −210 to −171 bp may also regulate the PAI-1 promoter activity. Between −566 and −561 bp a canonical E box (CACGTG) is present and between −194 and −189 bp an E box-like sequence (CACGTA) is present. As HGF slightly increased the E box mutated reporter activity, the distal E box site is likely active as another HGF response site with the mutated long construct. Indeed, USF1 binding to a consensus E box motif at nucleotides −566 to −561 bp in the PAI-1 gene was required for transforming growth factor (TGF)-β1-induced transcription of a PAI-1 promoter-driven luciferase reporter in human epidermal keratinocytes.23 As the longest construct that contains the region from −829 to +36 bp showed a modest increase in luciferase activity by HGF, additional DNA elements could be present outside of this region. Our data strongly suggested that USF1 and USF2 are involved in transcriptional regulation of PAI-1 expression through a functional E box at −158 to −153 bp. USFs control hepatic production of metabolic enzymes such as fatty acid synthetase.24 Thus, HGF mediated PAI-1 production may link thrombosis and metabolic derangements.
Co-transfection of cells with USF1 expression vector increased PAI-1 promoter activity, and increased PAI-1 promoter activity induced by HGF was attenuated by co-transfections with SREBP-1a and -1c expression vectors. Increased PAI-1 promoter activity induced by USF1 expression was attenuated by SREBP-1a. SREBPs belong to the same basic-helix-loop-helix family of transcriptional factors as do USFs. They can bind to an E box as well as to a sterol regulatory element.25 The results in this study suggest that competition between USFs and SREBPs for E box binding may account for the inhibitory effects of SREBPs on PAI-1 expression. Lack of inhibition by SREBP-2 is consistent with the idea that SREBP-1a and -1c are active for E box, whereas SREBP-2 is inactive.25
Because HGF phosphorylates USFs through MAPK and tyrosine kinase pathways and increased HGF is associated with insulin resistance, HGF may be a determinant of increased PAI-1 expression and evolution of atherothrombosis associated with metabolic derangements. Accordingly, targeting the HGF signaling pathway may ameliorate the prothrombotic risk and attenuate evolution of atherothrombosis.
The authors thank Drs R.G. Roeder and H. Shimano for providing the vectors encoding for human USF and SREBP. The authors greatly appreciate the technical assistance of Miwako Fujii and secretarial support of Lori Dales.
Sources of Funding
This work was supported in part by grants-in-aid for scientific research from the Ministry of Education, Science, Sport, and Culture of Japan. D.J. was supported by Japan Society for the Promotion of Science.
Original received December 28, 2005; final version accepted June 22, 2006.
Dandona P, Aljada A, Chaudhuri A, Mohanty P, Garg R. Metabolic syndrome: a comprehensive perspective based on interactions between obesity, diabetes, and inflammation. Circulation. 2005; 111: 1448–1454.
Heeschen C, Dimmeler S, Hamm CW, Boersma E, Zeiher AM, Simoons ML; CAPTURE (c7E3 Anti-Platelet Therapy in Unstable REfractory angina) Investigators. Prognostic significance of angiogenic growth factor serum levels in patients with acute coronary syndromes. Circulation. 2003; 107: 524–530.
Wojta J, Nakamura T, Fabry A, Hufnagl P, Beckmann R, McGrath K, Binder BR. Hepatocyte growth factor stimulates expression of plasminogen activator inhibitor type 1 and tissue factor in HepG2 cells. Blood. 1994; 84: 151–157.
Marchesini G, Brizi M, Bianchi G, Tomassetti S, Bugianesi E, Lenzi M, McCullough AJ, Natale S, Forlani G, Melchionda N. Nonalcoholic fatty liver disease: a feature of the metabolic syndrome. Diabetes. 2001; 50: 1844–1850.
Alessi MC, Bastelica D, Mavri A, Morange P, Berthet B, Grino M, Juhan-Vague I. Plasma PAI-1 levels are more strongly related to liver steatosis than to adipose tissue accumulation. Arterioscler Thromb Vasc Biol. 2003; 23: 1262–1268.
Dong J, Fujii S, Li H, Nakabayashi H, Sakai M, Nishi S, Goto D, Furumoto T, Imagawa S, Zaman T, Kitabatake A. Interleukin-6 and mevastatin regulate plasminogen activator inhibitor-1 through CCAAT/enhancer-binding proteinδ. Arterioscler Thromb Vasc Biol. 2005; 25: 1078–1084.
Sayasith K, Lussier JG, Sirois J. Role of upstream stimulatory factor phosphorylation in the regulation of the prostaglandin G/H synthase-2 promoter in granulosa cells. J Biol Chem. 2005; 280: 28885–28893.
Amemiya-Kudo M, Shimano H, Hasty AH, Yahagi N, Yoshikawa T, Matsuzaka T, Okazaki H, Tamura Y, Iizuka Y, Ohashi K, Osuga J, Harada K, Gotoda T, Sato R, Kimura S, Ishibashi S, Yamada N. Transcriptional activities of nuclear SREBP-1a, -1c, and -2 to different target promoters of lipogenic and cholesterogenic genes. J Lipid Res. 2002; 43: 1220–1235.
Kaneko T, Fujii S, Matsumoto A, Goto D, Ishimori N, Watano K, Furumoto T, Sugawara T, Sobel BE, Kitabatake A. Induction of plasminogen activator inhibitor-1 in endothelial cells by basic fibroblast growth factor and its modulation by fibric acid. Arterioscler Thromb Vasc Biol. 2002; 22: 855–860.
Cheung E, Mayr P, Coda-Zabetta F, Woodman PG, Boam DS. DNA-binding activity of the transcription factor upstream stimulatory factor 1 (USF-1) is regulated by cyclin-dependent phosphorylation. Biochem J. 1999; 344: 145–152.
Sirito M, Lin Q, Maity T, Sawadogo M. Ubiquitous expression of the 43- and 44-kDa forms of transcription factor USF in mammalian cells. Nucleic Acids Res. 1994; 22: 427–433.
Casado M, Vallet VS, Kahn A, Vaulont S. Essential role in vivo of upstream stimulatory factors for a normal dietary response of the fatty acid synthase gene in the liver. J Biol Chem. 1999; 274: 2009–2013.