Identification and Localization of a Fatty Acid Response Region in the Human Plasminogen Activator Inhibitor-1 Gene
Abstract—The increased expression of plasminogen activator inhibitor type-1 (PAI-1) is associated with increased concentrations of fatty acids in blood and may accelerate atherogenesis in diabetes. The present study was designed to define mechanisms by which nonesterified (free) fatty acids (FFAs) augment the expression of PAI-1. FFAs increased PAI-1 protein and mRNA expression by HepG2 cells. To identify potential regulatory elements, we constructed chimeric genes by fusing 1313, 853, 610, or 328 bp of human PAI-1 5′-flanking DNA to a luciferase reporter (PAI-LUC). A 2-fold increase in luciferase activity was seen when cells were transfected with PAI-LUC 1313, 863, or 610 and exposed to FFAs. No response to FFAs was seen with PAI-LUC 328 and after deletion of a 72-bp (−599 to −528) fragment from PAI-LUC 1313. This 72-bp fragment conferred FFA responsiveness to a different (simian virus 40) promoter. Two footprinted regions were demonstrated by DNase I analysis. Gel mobility shift assays indicated specific binding of extracted proteins to an FFA response element: 5′-TG(G/C)1–2CTG-3′. This sequence is repeated 4 times and is similar to an Sp1-binding site. Sp1 consensus oligonucleotides inhibited binding of extracted proteins to the regulatory elements. Accordingly, FFA-induced increased expression of PAI-1 in HepG2 cells is mediated by the binding of a transcription factor or factors to the repeated fatty acid response element, 5′-TG(G/C)1–2CTG-3′, that is highly homologous to an Sp1-binding site.
- Received July 19, 2000.
- Accepted September 5, 2000.
Plasminogen activator inhibitor type 1 (PAI-1), the primary physiological inhibitor of plasminogen activators, is increased in blood in patients with diabetes mellitus.1 Altered lipid metabolism is associated with increased concentrations of triglycerides and nonesterified (free) fatty acids (FFAs) in these subjects.2 We have found that the infusion of glucose and emulsified triglycerides (in combination with heparin to increase FFA concentrations) into healthy human subjects increases the concentration of PAI-1 in blood.3 Further, the exposure of cells to triglycerides increases elaboration of PAI-1 protein secondary to increased transcription.4 5 6 In previous work, we found that the effect of triglycerides on expression of PAI-1 appears to be mediated, at least in part, by their constituents: FFAs.7
Both FFAs and their metabolites, particularly long-chain acyl-CoA, have been shown to regulate the expression of genes through binding to or modification of transcription factors.8 9 10 11 We postulated that the induction of PAI-1 by FFAs is secondary to a direct or an indirect effect of FFAs on the expression and/or activation of a transcription factor or factors that bind to a cis-acting element in the 5′ flanking region of the PAI-1 gene. The present study was designed to identify the cis-acting element or elements.
Human hepatoma (HepG2) cells were obtained from American Type Culture Collection and grown in minimal essential media (MEM; GIBCO BRL) with 10% NuSerum (GIBCO BRL). Experiments were performed in DMEM with Ham’s nutrient mixture F12 (DME-F12; GIBCO BRL).
The sodium salts of FFAs (Sigma Chemical Co) were dissolved in water at 37°C and added dropwise to 1% fatty acid–free bovine serum albumin (FAF-BSA; Sigma Chemical Co) in DME-F12. We determined that 3% BSA in DME-F12 contains 0.45 mmol/L FFAs (Wako NEFA C kits; Biochemical Diagnostics). All solutions were adjusted to pH 7.4 and filtered through a 0.22-μm filter.
Eighty percent confluent HepG2 cells were preincubated in serum-free DME-F12 for 4 to 8 hours. Subsequently, cells were exposed to DME-F12 with FAF-BSA, DME-F12 with FAF-BSA to which selected fatty acids were added, or DME-F12 with BSA. Conditioned media were collected after selected intervals and assayed for PAI-1 by ELISA (Tintelize; Biopool). Total cellular RNA was isolated conventionally with TRIzol Reagent (GIBCO BRL). Northern blotting was performed as previously described with 10 μg RNA.12
Construction of PAI-1 Promoter-Luciferase Reporter Plasmids
A 1387-bp KpnI fragment containing 1313 bp of 5′-flanking (−1313) and 74 bp of the untranslated first exon (+74) of human PAI-1 DNA was isolated from the plasmid PAI-CAT 1313,13 inserted into the unique KpnI site of the promoter-less luciferase reporter plasmid pGL3-Basic (Promega), and referred to as PAI-LUC 1313. A 927-bp XhoI fragment (from −853 to +74 in human PAI-1 DNA) was inserted into the XhoI site of pGL3-Basic and referred to as PAI- LUC 853. Polymerase chain reaction was used to amplify a 684-bp fragment that encompassed −610 to +74 (5′ primer, 5′-cta ggt acc aga cca aga gtc ctc tgt tg-3′; 3′ primer, 5′-aat gga tcc gaa ttc agc tgc tgg agg-3′). This fragment was inserted into pGL3-Basic and referred to as PAI-LUC 610. A 402-bp HindIII fragment containing −328 to +74 bp of the PAI-1 gene was inserted into pGL3-Basic and referred to as PAI-LUC 328. A 72-bp ApaI fragment (−599 to −528) containing a potential fatty acid response element14 was deleted from PAI-LUC 1313 to form PAI-LUC 1313−72. To generate PAI-LUC 328+72, the 72-bp ApaI fragment was inserted upstream of the minimal promoter (PAI-LUC 328) in the sense orientation. The 72+pGL3-Promoter was generated by inserting the 72-bp KpnI/XhoI fragment upstream of the simian virus (SV)40 promoter in pGL3-Promoter (Promega).
Transient transfection of HepG2 cells was performed according to the calcium-phosphate precipitation method as previously described.13 Transfections were performed with 12 μg reporter plasmid and 600 ng pRL-cytomegalovirus (CMV), a plasmid with Renilla luciferase gene downstream of the CMV promoter to control for transfection efficiency. Luciferase activity was detected in cell extracts with the Passive Lysis Buffer in the Dual-Luciferase Reporter Assay System (Promega). Luciferase activities were determined with the Dual-Luciferase Reporter Assay with a DLReady luminometer (Turner Designs Instrument).
DNase I footprinting was performed with −618 to −487 of the PAI-1 gene in accordance with the Maxam method.15 Protein extract (100 μg) was used in a 50-μL reaction that contained 10 mmol/L Tris, pH 7.5, 50 mmol/L NaCl, 1 mmol/L EDTA, 1% Ficoll, and 1 μg poly(dI/dC)·poly(dI/dC).
Electrophoretic Mobility Shift Assays
Proteins were extracted from HepG2 cells exposed to control media (DME-F12 with 1% FAF-BSA) and to 0.75 mmol/L capric acid in DME-F12 with 1% FAF-BSA according to the freeze-thaw method described by Ladias et al.16
Oligonucleotides of PAI-1 5′ flanking DNA and Sp1 consensus sequences17 that were used in electrophoretic mobility shift assays (EMSAs) were synthesized at Life Technology (GIBCO BRL) and include 29 bp, 5′-gcatgccctgtggctgttgggctgggccc-3′ and 5′-gggcccagcccaacagccacagggcatgc-3′; 15 bp, 5′-tggctgttgggctgg-3′ and 5′-ccagcccaacagcca-3′; 9 bp, 5′-tggctgttg-3′ and 5′-caacagcca-3′; 23 bp, 5′-cacgtggctggctgcatgccctg-3′ and 5′-cagggcatgcagcca-gccacgtg-3′; 18 bp, 5′-cacgtggctggctgcatg-3′ and 5′-catgcagcca-gccacgtg-3′; 13 bp, 5′-cacgtggctggct-3′ and 5′-agccagccacgtg-3′; 10 bp, 5′-gcatgccctg-3′ and 5′-cagggcatgc-3′; C13, 5′-tggctgctggctg-3′ and 5′-cagccagcagcca-3′; M13, 5′-taactgctaactg-3′ and 5′-cagttagcagtta-3′; 30 bp, 5′-aagtcctagacagacaaaacctagacaatc-3′ and 5′-gattgtctaggttttgtctgtctaggactt-3′; Sp1, 5′-acaattgggctgggcctaat-3′ and 5′-attaggcccagcccaattgt-3′ and 5′-attcgatcggggcggggcgagc-3′ and 5′-gctcgccccgccccgatcgaat-3′ (Promega); and 24 bp, 5′-tcacagaacatgtctctatcgtaa-3′ and 5′-ttacgatagagagatgttcagtga-3′ (control).
Double-strand oligonucleotides were end-labeled with (γ-32P)-ATP and T4 polynucleotide kinase. Unincorporated nucleotides were removed with a NucTrap column (Stratagene). EMSAs were performed as previously described.18
Results are given as mean±SEM. Differences between 2 groups were identified with a Student’s t test. For multiple groups, 1-way ANOVA and Student-Newman-Keuls tests were used to identify differences. Significance was defined as P<0.05.
FFAs and Accumulation of PAI-1 in Conditioned Media
Increased PAI-1 was seen in media conditioned with HepG2 cells after exposure to FFAs. Compared with FAF-BSA, 3% BSA containing a physiological mixture of FFAs increased 24-hour accumulation of PAI-1 by 2.2±0.8-fold (n=6, P<0.01). To confirm that effects with BSA were mediated by FFAs, cells were exposed to selected FFAs, including capric acid, linoleic acid, and lauric acid. Similar increases were seen in the accumulation of PAI-1 in 24-hour–conditioned media (capric acid by 3.0±0.1-fold; linoleic acid by 2.6±0.1-fold; and lauric acid by 2.5±0.2-fold, n=6 for each, P<0.005). Increases were concentration dependent (Figure 1⇓).
FFAs and Expression of PAI-1 mRNA
Exposure to 0.75 mmol/L capric acid led to a 2±0.02-fold increase in 3.2-kb PAI-1 mRNA and a 2.8±0.1-fold increase in 2.2-kb PAI-1 mRNA (n=3, P<0.001). Maximal increases were seen 16 to 48 hours after exposure.
FFAs and Luciferase Activity
HepG2 cells were transfected with PAI-LUC reporter constructs (Figure 2⇓). A 1.7-fold increase in luciferase activity (corrected for transfection efficiency) was seen in extracts from cells transfected with PAI-LUC 1313, 853, or 610 after exposure to 0.75 mmol/L albumin-bound FFAs (n=6 for each). Comparable results were seen with capric acid and with 3% BSA binding a physiological mixture of FFAs (fold induction with 3% BSA, 1.36±0.02 for PAI-LUC 1313 and 1.35±0.03 for PAI-LUC 610, n=6 for each, P<0.001).
After the deletion of a 72-bp ApaI fragment (−599 to −528, PAI-LUC 1313−72), neither 3% BSA nor capric acid augmented luciferase activity. FFA-induced increased luciferase activity was seen when the 72-bp segment was inserted upstream of the minimal promoter (PAI-LUC 328+72, 1.43±0.04-fold increase with 3% BSA, n=6, P<0.001). Increased FFA-induced luciferase activity was also seen when the 72-bp fragment was inserted upstream of the SV40 promoter in pGL3-Promoter (1.29±0.01- fold, n=8, P<0.001). Accordingly, the 72-bp segment, from −599 to −528 of the PAI-1 gene, contains the fatty acid response region.
Interaction Between Extracted Proteins and Fatty Acid Response Region
DNase I footprinting was performed with a 130-bp end-labeled fragment spanning −618 to −488 in the PAI-1 promoter region. This segment contains the 72-bp region with the implicated fatty acid response element in addition to a 20-bp sequence upstream and a 40-bp sequence downstream. The fragment was exposed to extracts from cells treated with 0.75 mmol/L capric acid. Two regions (referred to as A and B) in the end-labeled 130-bp fragment were protected from DNase I digestion (Figure 3⇓). Footprinted site A corresponds to nucleotides −524 to −558. Footprinted site B corresponds to nucleotides −559 to −590.
Interaction of DNA With Proteins Extracted From HepG2 Cells
One fragment that encompasses the entire footprinted A region, 2 overlapping fragments (29 and 23 bp) that encompass the footprinted A region, and 1 fragment that encompasses the footprinted B region (30 bp) were used for gel mobility shift analysis. These fragments exhibited specific binding to proteins extracted from HepG2 cells (n=5 for each; Figure 4⇓). Results with the fragment that encompasses the entire footprinted region A were similar to those seen with the 29-bp fragment (data not shown). Although 3 DNA/protein complexes were identified with the selected fragments, only binding of complex I was consistently inhibited with self “cold” (non-labeled). No inhibition was seen when a 24-bp unrelated fragment (100-fold excess) was used as competitor (Figure 4⇓).
FFAs and Binding of Proteins Extracted From HepG2 Cells to Fatty Acid Response Region
Proteins extracted from cells exposed to control conditions (FAF-BSA) and to FFAs were exposed to oligonucleotides in EMSAs. Induction of complex I was seen when the 29- or 30-bp probe was exposed to proteins from cells treated with FFAs. Increased binding was seen 5 to 120 minutes after exposure to FFAs (n=3 for each).
Identification of Fatty Acid Response Element
DNA sequence analysis of the fatty acid responsive region identified 4 repeats of the sequence 5′-TG(G/C)1–2CTG-3′ in the 23- and 29-bp fragments. The EMSAs demonstrated that the fatty acid responsive element is common to the 29- and 23-bp fragments. Progressively smaller components of the 29- and 23-bp fragments were used to delineate further the fatty acid response element. Complex I was seen with each fragment, including those with a single copy of the implicated fatty acid response element (Figure 5⇓). In addition, EMSAs were performed with a 13-bp oligonucleotide that contained the sequence 5′-TGGCTG C TGGCTG-3′. This oligonucleotide contains 2 repeats of the implicated fatty acid response element with 1 spacer (C). The same shifted pattern was seen when the oligonucleotide was incubated with proteins extracted from HepG2 cells. Mutation of “GG” to “AA” abolished the binding when the probe 5′-TAACTG C TAACTG-3′ was used in the EMSAs.
Involvement of a Transcription Factor That Binds to Sp1 Consensus Sites
Footprinted region A contains 4 repeats of the implicated fatty acid response element. The sequence is similar to the Sp1 binding sequence (T/G)(G/A)GGC(T/G)G(G/A)(G/A)(C/T). In addition, footprinted region B contains a 10-bp segment that has 70% identity with the Sp1 consensus site. Thus, EMSAs were performed with Sp1 consensus oligonucleotides and demonstrated a DNA-protein complex with an electrophoretic mobility pattern similar to that seen with complex I (Figure 6⇓). Binding of extracted protein to complex I with the 29- or 23-bp fragment was inhibited with a 100-fold molar excess of unlabeled Sp1 consensus DNA (Figure 6⇓). Inhibition was seen also when extracted proteins were exposed to unlabeled 29- or 23-bp fragments before exposure to radiolabeled Sp1 consensus oligonucleotides.
The 30-bp fragment inhibited binding of proteins to complex I when competition EMSAs were performed with the 29-bp, 23-bp, or Sp1 consensus (Figure 6⇑). Similarly, unlabeled 29-bp, 23-bp, and Sp1 consensus oligonucleotides inhibited the binding of proteins to the radiolabeled 30-bp fragment. Thus, the results of the competition EMSAs suggest that a protein that can bind to Sp1 consensus DNA increases transcription of the PAI-1 gene after exposure of the HepG2 cells to FFAs.
Patients with type II diabetes have increased concentrations of FFAs and PAI-1 in their blood.1 19 The increased expression of PAI-1 is likely to accelerate atherogenesis and to contribute to an increased incidence of acute myocardial infarction.20 21 22 Increased PAI-1 in blood predisposes to more persistent and more exuberant thrombosis by impairing the fibrinolytic response. Accordingly, vessel walls may be exposed to increased concentrations of clot-associated mitogens for prolonged intervals. Identification of the mechanisms responsible for increased expression of PAI-1 should facilitate the development of novel therapies with the aim of retarding the progression of atherosclerosis and preventing myocardial infarction.
In the present study, we found that a physiological mixture of fatty acids associated with BSA as well as 2 medium-chain length fatty acids (capric acid and lauric acid) and a long-chain polyunsaturated fatty acid (linoleic acid) increase the expression of PAI-1 in HepG2 cells. Our results are consistent with those of others that demonstrate increased expression of PAI-1 by FFAs of varying chain length and saturation.4 23 24 Thus, the effect of fatty acids on the expression of PAI-1 appears to be related to the prevailing concentration of diverse FFAs rather than to the concentration of a specific fatty acid.
We used chimeric genes that contain firefly luciferase as a reporter downstream of selected segments of 5′ flanking DNA from the human PAI-1 gene to identify the mechanism by which fatty acids augment the expression of PAI-1. Deletion analysis of the PAI-1 promoter region indicated that the fatty acid response region is located between −599 and −528 in the PAI-1 5′ flanking region. The implicated fatty acid response element is distinct from the VLDL response element identified by Erikssen et al.6 These results suggest that FFAs augment the expression of PAI-1 independent of effects mediated by triglycerides or lipoproteins.
DNA sequence analysis demonstrated that the implicated fatty acid response sequence was repeated 4 times in the footprinted region A. The footprinted region B contains a sequence “TGTCTG” in the complementary strand that is 83% identical to the implicated fatty acid response element. Sp1 recognition sequences are functional in either orientation17 Thus, despite the asymmetry of Sp1-binding sites, competition between the 30-bp fragment and the 29- or 23-bp fragment would be expected. Further, the increased binding of proteins to the fatty acid response element seen within 5 minutes after exposure of the cells to FFAs suggests activation rather than induction of a transcription factor. Accordingly, increased expression of PAI-1 after exposure of the HepG2 cells to FFAs is secondary to activation of a transcription factor that binds to a repeated element in the 5′ flanking region of the PAI-1 gene.
Three fatty acid response regions have been described previously.10 25 26 A cis-linked fatty acid response element was localized to a 140-bp (−80 to −220) region within the S14 proximal promoter.10 25 26 In the pyruvate kinase gene, a fatty acid response region was identified that overlaps the glucose/insulin response element. These results suggest that fatty acids inhibit the pyruvate kinase transcription by interfering with glucose and insulin induction of the gene.25 Specific binding of trans-acting proteins to the response regions was not defined in these studies. The third fatty acid responsive region was identified in the stearoyl-CoA desaturase gene 1 (SCD 1). Proteins from HepG2 cells and 3T3-L1 cells have been shown to bind specifically to a 60- bp region in the SCD1 promoter.26 Thus, the present results constitute the first report of a Sp1 consensus site that mediates the effects of fatty acids.
Sp1-like proteins are a family of proteins that contain 3 highly conserved C-terminal zinc finger domains and bind to GC-rich sequences.17 27 Consistent with our observation, Sp1-responsive promoters usually contain multiple recognition sites, although a single binding site appears to be sufficient for a promoter to be stimulated by Sp1.17
Previous studies have demonstrated that Sp1 sites contribute to transcriptional activation of the rat PAI-1 gene.28 Proteins extracted from HTC rat hepatoma cells bind to Sp1 and CTF/NF-1–like binding sites in the rat PAI-1 regulatory region.28 An Sp1-binding site identified in the rat PAI-1 gene corresponds to the binding sites identified in footprint A in the present study. The activation of Sp1 in a different region of the PAI-1 gene has been implicated in glucose-mediated increased expression of PAI-1.29 In studies with rat aortic smooth muscle cells, glucose has been shown to release a transcriptional repressor from Sp1 complexes. By contrast, we found that FFAs activate a transcription factor that binds to Sp1 consensus sites. Thus, the increased expression of PAI-1 seen in subjects with diabetes appears to be secondary, at least in part, to increased transcription of PAI-1 induced by Sp1-binding sites and mediated by increased concentrations of both glucose and FFAs.
In summary, we have found that fatty acids augment the expression of PAI-1 in a human hepatoma cell line (HepG2). FFAs activate a transcription factor that binds to the sequence 5′-TG(G/C)1–2CTG-3′ in the 5′ flanking DNA of the human PAI-1 gene. This sequence is nearly identical to the DNA-binding site for Sp1. Further characterization of the transcription factor responsible for FFA-induced augmented expression of PAI-1 and its interaction with other transcription factors may provide a therapeutic target for modification of the expression of PAI-1 in diabetic subjects.
This project was supported by a grant from the American Diabetes Association (Dr Schneider).
McGill JB, Schneider DJ, Arfken CL, Lucore CL, Sobel BE. Factors responsible for impaired fibrinolysis in obese subjects and NIDDM patients. Diabetes. 1994;43:104–109.
Auwerx J, Bouillon R, Collen D, Geboers J. Tissue-type plasminogen activator antigen and plasminogen activator inhibitior in diabetes mellitus. Arteriosclerosis. 1998;8:68–72.
Calles-Escandon J, Mirza SA, Sobel BE, Schneider DJ. Induction of hyperinsulinemia combined with hyperglycemia and hypertriglyceridemia increases plasminogen activator inhibitor 1 in blood in normal human subjects. Diabetes. 1998;47:290–293.
Banfi C, Risé P, Mussoni l., Galli C, Tremoli E. Linoleic acid enhances the secretion of plasminogen activator inhibitor type 1 by HepG2 cells. J Lipid Res. 1997;38:860–869.
Sironi L, Mussoni l., Prati L, Baldassarre D, Camera M, Banfi C, Tremoli E. Plasminogen activator inhibitor type-1 synthesis and mRNA expression in HepG2 cells are regulated by VLDL. Arterioscler Thromb Vasc Biol. 1996;16:89–96.
Eriksson P, Nilsson L, Karpe F, Hamsten A. Very-low-density lipoprotein response element in the promoter region of the human plasminogen activator inhibitor-1 gene implicated in the impaired fibrinolysis of hypertriglyceridemia. Arterioscler Thromb Vasc Biol. 1998;18:20–26.
Amri EZ, Ailhaud G, Grimaldi PA. Fatty acids as signal transducing molecules: involvement in the differentiation of preadipose to adipose cells. J Lipid Res. 1994;35:930–937.
Jump DB, Clarke SD, MacDougald O, Thelen A. Polyunsaturated fatty acids inhibit S14 gene transcription in rat liver and cultured hepatocytes. Proc Natl Acad Sci U S A. 1993;90:8454–8458.
Keller H, Dreyer C, Medin J, Mahfoudi A, Ozato K, Wahli W. Fatty acids and retinoids control lipid metabolism through activation of peroxisome proliferator-activated receptor-retinoid X receptor heterodimers. Proc Natl Acad Sci U S A. 1993;90:2160–2164.
Schneider DJ, Sobel BE. Augmentation of synthesis of plasminogen activator inhibitor type 1 by insulin and insulin-like growth factor type I: implications for vascular disease in hyperinsulinemic states [published erratum appears in Proc Natl Acad Sci U S A. 1992;89:1148]. Proc Natl Acad Sci U S A. 1991;88:9959–9963.
Westerhausen DRJ, Hopkins WE, Billadello JJ. Multiple transforming growth factor-beta-inducible elements regulate expression of the plasminogen activator inhibitor type-1 gene in Hep G2 cells. J Biol Chem. 1991;266:1092–1100.
Ladias JA, Hadzopoulou-Cladaras M, Kardassis D, Cardot P, Cheng J, Zannis V, Cladaras C. Transcriptional regulation of human apolipoprotein genes apoB, apoCIII, and apoAII by members of the steroid hormone receptor superfamily HNF-4, ARP-1, EAR-2, and EAR-3. J Biol Chem. 1992;267:15849–15860.
Fried M, Crothers DM. Equilibria and kinetics of lac repressor-operator interactions by polyacrylamide gel electrophoresis. Nucleic Acids Res. 1981;9:6505–6525.
Loskutoff DJ, van Aken BE, Seiffert D. Abnormalities in the fibrinolytic system of the vascular wall associated with atherosclerosis. Ann N Y Acad Sci. 1995;748:177–83.
Hagen G, Muller S, Beato M, Suske G. Cloning by recognition site screening of two novel GT box binding proteins: a family of Sp1 related genes. Nucleic Acids Res. 1992;20:5519–5525.
Johnson MR, Bruzdzinski CJ, Winograd SS, Gelehrter TD. Regulatory sequences and protein-binding sites involved in the expression of the rat plasminogen activator inhibitor-1 gene. J Biol Chem. 1992;267:12202–12210.
Chen YQ, Su M, Walia RR, Hao Q, Covington JW, Vaughan DE. Sp1 sites mediate activation of the plasminogen activator inhibitor-1 promoter by glucose in vascular smooth muscle cells. J Biol Chem. 1998;273:8225–8231.