Genotype-Specific Transcriptional Regulation of PAI-1 Expression by Hypertriglyceridemic VLDL and Lp(a) in Cultured Human Endothelial Cells
Abstract The hypothesized relationships between plasminogen activator inhibitor (PAI-1) genotypes, PAI-1 levels, and their potential regulation by hypertriglyceridemic (HTG) very low density lipoprotein (VLDL) and lipoprotein(a) [Lp(a)] was examined in a PAI-1 genotyped human umbilical vein endothelial cell (HUVEC) culture model system. Individual human umbilical veins were used to obtain cultured ECs and were genotyped for PAI-1 by using the HindIII restriction fragment length polymorphism (RFLP) as a marker for genetic variation. Digested genomic DNA, examined by Southern blot analysis and probed with an [α-32P]dCTP–labeled 2.2-kb PAI-1 cDNA, yielded three RFLPs designated 1/1 (22-kb band only), 1/2 (22- plus 18-kb bands), and 2/2 (18-kb band only). Individual PAI-1 genotyped HUVEC cultures were incubated in the absence or presence of HTG-VLDL (0 to 50 μg/mL) or Lp(a) (0 to 50 μg/mL) at 37°C for various times (4 to 24 hours), followed by analyses of PAI-1 antigen (by ELISA) and mRNA (by ribonuclease protection assay) levels, EC surface–localized plasmin generation assays, and nuclear run-on transcription assays. Secreted PAI-1 antigen levels were increased ≈2- to 3-fold by HTG-VLDL and ≈1.6 to 2-fold by Lp(a); mRNA levels were increased ≈3- to 4.5-fold by HTG-VLDL and ≈2.5- to 3.2-fold by Lp(a) compared with medium-incubated controls, primarily in the 2/2 PAI-1 genotype HUVEC cultures. Increases in PAI-1 mRNA induced by HTG-VLDL or Lp(a) could be abolished by coincubation with actinomycin D (2×10−6 mol/mL) or puromycin (1 μg/mL). In addition, nuclear transcription run-on assays typically demonstrated that HTG-VLDL increased PAI-1 gene transcription rates by ≈5- to 6-fold and ≈4- to 5-fold, respectively, primarily in the 2/2 PAI-1 genotype HUVEC cultures compared with 1/1 PAI-1 genotype HUVEC cultures or medium-incubated controls. The positive control interleukin-1 increased both 2/2 and 1/1 PAI-1 mRNA levels by ≈5- to 6-fold. Increased PAI-1 antigen and mRNA expression were associated with a concomitant 50% to 60% decrease in plasmin generation. These combined results demonstrate the genotype-specific regulation of PAI-1 expression by HTG-VLDL and Lp(a) and further indicate that these risk factor–associated components regulate PAI-1 gene expression at the transcriptional level in cultured HUVECs. Results from these studies further suggest that individuals with this responsive 2/2 PAI-1 genotype may reflect the additional inherent potential for later HTG-VLDL- or Lp(a)-induced fibrinolytic dysfunction, resulting in the early initiation of thrombosis, atherogenesis, and coronary artery disease.
- Received June 10, 1996.
- Accepted May 29, 1997.
Normal hemostasis and vascular patency are maintained by a balance between coagulation and fibrinolysis. ECs play an important role in this process by producing coagulation proteins that promote clot formation, fibrinolytic proteins (TPA and UPA as well as their inhibitor, PAI-1) that facilitate clot lysis, and localization of these proteins to the cell surface via specific receptors or binding proteins.1 2 Fibrinolytic dysfunction, in particular elevated plasma levels of PAI-1, has been associated with thromboembolic disease and may represent a risk factor for CAD and MI.3 4 5 Well-established risk factors for CAD and MI such as smoking, obesity, hyperlipoproteinemia, and non–insulin-dependent diabetes (type II diabetes) are commonly associated with impaired fibrinolysis and may be due in part to increased plasma PAI-1 levels.6 7 8 9 Because PAI-1 is generally considered to be a major regulator of fibrinolysis through its interaction with plasminogen activators (TPA and UPA),10 11 it is conceivable that elevated plasma PAI-1 levels in certain hyperlipoproteinemias may partially explain the increased thrombotic risk and CAD prevalence, which are otherwise unexplained by conventional risk factors.9 12
Variations in plasma PAI-1 levels have been related to variations in the PAI-1 gene. Three polymorphic variations in the human PAI-1 gene have been reported: a single nucleotide insertion/deletion (4G/5G) polymorphism in the promoter region; an allelic variation at a (CA)n dinucleotide repeat polymorphism; and an HindIII RFLP due to a base change at the 3′UTR of the PAI-1 gene.13 14 15 16 A genotype-specific association has been found in vivo between the HindIII RFLPs, PAI-1 levels, and HTG-VLDLs,16 which has led to the hypothesis that a relationship may exist between a responsive PAI-1 genotype, PAI-1 levels, and the potential regulation by VLDL.16 In another study, however, this relationship was not confirmed.17 The studies described herein have demonstrated that this relationship exists in vitro through the use of a PAI-1 genotyped (HindIII RFLP) human EC culture model system and that both HTG-VLDLs and Lp(a) regulate PAI-1 expression in a genotype-specific manner.
Human Glu-Pmg was obtained from Enzyme Research Products, Inc; TPA from American Diagnostica Inc; collagenase (type I, CLS) from Boehringer Mannheim Biochemicals; fetal bovine serum from Intergen Corp; heparin (porcine intestinal mucosa) and BSA from Sigma Chemical Co; [α-32P]dCTP, [α-32P]UTP (3000 Ci/mmol), and [125I]NaI (specific activity, 14.0 mCi/μg) from Amersham Corp; Iodo-Beads from Pierce Chemical Co; Sephadex G-25 columns (PD-10) from Pharmacia; Aprotinin (Trasylol) from Mobay Corp; DiI-Ac-LDL from Biomedical Technologies, Inc; medium 199, TRIzol reagent (total RNA isolation reagent), and UPA and PAI-1 ELISA kits (TintElize) from BioPool; a random-primer kit (Primer II) and Quick Hyb solution from Stratagene; proteinase K, HindIII, the RPA kit, HybSpeed RPA, TRIPLEscript promoter cloning vector, transcription kit, MAXIscript, and pTRI-GAPDH–human plasmid from Ambion; and DNA molecular weight marker XV (2.4- to 48.5-kb range) from Boehringer Mannheim. Frozen bovine hypothalamuses from which EC growth factor was isolated18 were obtained from Pel-Freeze Inc. HTG-VLDL19 and Lp(a)20 21 were isolated and characterized as described previously.
HUVECs were obtained from freshly discarded umbilical veins by mild collagenase treatment.1 2 19 ECs were seeded into plastic Petri dishes (960 mm2) coated with human fibronectin and grown to confluence in complete culture medium, as we have previously described in detail.1 2 19 All experiments were carried out with individual vessel-derived, first-passage, postconfluent (2 to 3 days after reaching their stable confluence density, ie, ≈8.9 to 9.2×105 cells per cm2) cultured HUVECs. Cells were routinely counted by using phase-contrast microscopy and a 0.5×0.5-mm counting reticle.
PAI-1 antigen/mRNA analyses were carried out with postconfluent HUVEC cultures grown in 24-well multiwell plates (200 mm2 per well) representing each of the three individual HindIII PAI-1 genotypes: 1/1, 1/2, and 2/2 (see below). Cultures were incubated (0 to 24 hours) in serum-free culture medium (500 μL per well of the same composition as as the complete culture medium, except that FBS was replaced by 0.25% BSA) in the absence or presence of HTG-VLDL (0 to 50 μg/mL) or Lp(a) (0 to 50 μg/mL), and each individual genotyped culture (in triplicate) was then analyzed for its secreted PAI-1 antigen and coincident mRNA levels as described below; UPA antigen levels were also analyzed as the control protein. Experiments with inhibitors were also carried out in genotyped HUVEC cultures that were preincubated for 30 minutes with actinomycin D (2×10−6 mol/L) or puromycin (1 μg/mL) followed by incubation in the absence or presence of the lipoproteins for an additional 8 hours at 37°C in the presence of the respective inhibitor.
Nuclear transcription run-on assays were carried with postconfluent cultured HUVECs grown in T-75 (7500-mm2) flasks representing the 1/1 and 2/2 HindIII PAI-1 genotypes. Cultures were incubated in serum-free medium 199 containing 0.5% BSA in the absence or presence of HTG-VLDL, Lp(a), or IL-1 (positive control) for 8 hours prior to isolation of the nuclei as described below.
Individual genotyped cell cultures were routinely characterized as ECs and their purity established by their uniform uptake of the EC-specific fluorescent probe DiI-Ac-LDL22 and their typical monolayer “cobblestone.” tight-packing growth morphology.22 23 24 25 26 27 28 29 Only individual cultures with >95% identifiable ECs were used in these experiments.
Genomic DNA was isolated from pieces of fresh human umbilical cords according to Maniatis et al.30 In brief, pieces of tissue (300 to 500 mg) were incubated in digestion buffer containing proteinase K (100 μg/mL) in the presence of 0.1 mol/L EDTA and 0.5% SDS for 16 hours at 55°C. Digested samples were then extracted with phenol saturated with 0.5 mol/L Tris-HCl, pH 8.0, and the DNA recovered by ethanol precipitation. DNA resuspended in 0.01 mol/L Tris-HCl, pH 8.0, and 0.001 mol/L EDTA was then incubated with 1 μg/mL DNase-free RNase for 30 minutes at 37°C followed by extraction with phenol/chloroform/isoamyl alcohol and ethanol precipitation. The DNA was washed once with 75% ethanol and resuspended in 0.010 mol/L Tris-HCl, pH 7.4. Absorbance ratios at 260 and 280 nm (A260 to A280) of DNA samples were >1.70.
Identification of HindIII RFLPs by Southern Blot Analysis
Klinger et al31 first described variations in the PAI-1 gene by using HindIII RFLPs as a marker for genetic variation. Genomic DNA (3 μg) was digested with HindIII (3 U/μg DNA) for 16 hours at 37°C and then electrophoresed overnight in a 0.4% agarose gel in 0.089 mol/L Tris, 0.089 mol/L boric acid, and 0.002 mol/L EDTA, pH 8.3, buffer,30 including appropriate DNA molecular weight markers (7.6- to 48.5-kb range). DNA digests were blotted onto nylon membrane by capillary transfer in 10× SSC (1× SSC is 0.15 mol/L NaCl and 0.015 mol/L sodium citrate buffer, pH 7.0) for 16 hours at room temperature and fixed to the nylon membrane. The immobilized DNA was hybridized with a 2.2-kb PAI-1 cDNA probe32 and radiolabeled with [α-32P]dCTP to a specific activity of ≈1×109 cpm/μg by using a random-primer kit according to the manufacturer’s instructions. Prehybridization and hybridization were carried out with Quick Hyb solution in a hybridization oven for 1 hour at 68°C. The filters were washed twice each in 2× SSC and 0.5% SDS for 30 minutes at room temperature, once in 1× SSC and 0.5% SDS for 30 minutes at 68°C, and once in 0.1× SSC and 0.1% SDS for 10 minutes at 68°C. The filters were exposed to Fuji film at −70°C for 24 hours with the use of Fisher Biotech L-Plus intensifying screens. From the Southern blot autoradiograms in conjunction with the DNA molecular weight markers, the presence of the 18-kb and/or 22-kb restriction fragment bands were used to identify the three different PAI-1 genotypes, designated in these studies as 1/1 (22-kb band only), 1/2 (22- plus 18-kb bands), and 2/2 (18-kb band only; containing the mutation site). All Southern blots were analyzed independently by three different investigators to identify and confirm each of the individual PAI-1 genotypes. These studies were carried out with a total of 94 different individual PAI-1 genotyped umbilical cord/HUVEC cultures (36 2/2’s, 29 1/2’s, and 29 1/1’s).
Lp(a) was isolated from the plasma of a single subject with two polymorphs of apo(a) by a procedure slightly modified from the one described previously.20 21 The purified Lp(a), in 33 mmol/L phosphate buffer, pH 7.4, containing 0.01% Na2EDTA and 0.01% NaN3, was sterilized by filtration (pore size, 0.45 μm) and stored at 4°C in sterile vials filled to the top to eliminate air space and minimize oxidation.
Measurement of PAI-1 and UPA Antigen Levels by ELISA
Total secreted PAI-1 and UPA antigen levels (free plus complex forms) were measured (in duplicate) in serum-free conditioned culture media by using a commercial PAI-1 and UPA ELISA (TintElize, BioPool) kit. Absorption was measured at 405 nm with a Dynatech plate reader and converted to each respective antigen concentration by the use of appropriate standards and BioLinx software (Dynatech). Total secreted PAI-1 and UPA antigen levels (24 hours) were calculated in nanograms per milliliter per well (≈1.8×105 cells per well). The data were finally expressed and compared as the PAI-1 or UPA antigen ratio obtained in the presence of HTG-VLDL or Lp(a) to its respective genotype control.
Total cytoplasmic RNA was isolated from the confluent monolayer of each individual genotyped EC culture (in duplicate) that had been incubated for varying times (0, 4, 8, 12, 18, and 24 hours), in the absence or presence of HTG-VLDL or Lp(a) (see above). Cell monolayers were washed twice in Dulbecco’s PBS, and total RNA was extracted by a single-step method using TRIzol Reagent according to the manufacturer’s instructions.
Measurement of PAI-1 mRNA Level by RPA
Relative changes in coincident PAI-1 mRNA levels after incubation (8 hours) in the absence or presence of HTG-VLDL or Lp(a) were measured by RPA. Total cytoplasmic RNA (0.5 to 1 μg) from each individual culture was hybridized (in duplicate) with a 212-base antisense probe of human PAI-1 (nucleotides 138 to 350). A 212-bp fragment of the PAI-1 cDNA was amplified by PCR by using a 24-mer upstream primer (5′ GTCTGCTGTGCACCATCCCCCATC 3′) identical to positions 138 to 161 and a 24-mer downstream primer (5′ TTGTCATCAATCTTGAATCCCATA 3′) complimentary to positions 327 to 350 of the human PAI-1 cDNA.32 This 212-bp fragment was then ligated into a TRIPLEscript promoter cloning vector. The PAI-1 probe was transcribed with [α-32P]UTP; 400 Ci/mmol, 10 mCi/mL), using an in vitro transcription kit (MAXIscript), to a specific activity of ≈1×109 cpm/μg to generate a 212-bp antisense PAI-1 probe. A 316-base antisense probe, purchased as a pTRI-GAPDH–human plasmid, was similarly transcribed and used to measure the internal control GAPDH.33 The RPA was carried out using the HybSpeed RPA kit according to the manufacturer’s instructions. RNase-protected, 32P-labeled RNA fragments were separated on a 5% denaturing polyacrylamide gel and dried, and the radioactivity content of each 32P-labeled, protected RNA fragment band was analyzed and quantified by phosphorimaging autoradiography. Radioactivity values were used to express PAI-1 mRNA to GAPDH mRNA ratios to reflect the relative changes in PAI-1 mRNA levels.
Nuclear Transcription Run-on Assays
Nuclear run-on assays were carried out essentially as described by Greenberg and Ziff.34 In brief, postconfluent HUVEC cultures in T-75 tissue culture flasks representing the 1/1 and 2/2 PAI-1 genotypes (one flask each in duplicate) were incubated in the absence or presence of HTG-VLDL (20 μg/mL), Lp(a) (50 μg/mL), or IL-1 (50 ng/mL) (positive control) at 37°C for 8 hours followed by a wash in Dulbecco’s PBS. Cells were scraped from the culture flasks and resuspended in NP-40 lysis buffer (0.001 mol/L EDTA, 0.0001 mol/L PMSF, 0.01 mol/L NaCl, 0.003 mol/L MgCl2, 10−6 mol/L antipain, 0.01 mol/L Tris-HCl, pH 7.4, and 0.5% vol/vol NP-40) and incubated on ice for 5 minutes. Intact nuclei were pelleted by centrifugation (1660g for 5 minutes at 4°C), washed once in NP-40 lysis buffer, resuspended in glycerol storage buffer (0.01 mol/L Tris-HCl, pH 8.3, 0.0001 mol/L EDTA, 0.001 mol/L DTT, and 40% glycerol) at 6×106 nuclei per 100 μL, and snap-frozen in liquid N2. Transcription reactions were carried out by mixing the nuclei (100 μL) with an equal volume (100 μL) of 2× reaction buffer (0.03 mol/L Tris-HCl, pH 8; 0.0005 mol/L MgCl2; 0.3 mol/L KCl; 25 U/mL placental RNAse; 0.01 mol/L creatine phosphate; 20 U/mL creatine phosphokinase; 0.001 mol/L each of ATP, CTP, and GTP; 0.0001 mol/L PMSF; and 0.0005 mol/L DTT) containing 100 μCi of [α-32P]UTP. After incubation at 28°C for 45 minutes, the reactions were terminated by adding a solution of 0.01 mol/L Tris-HCl, pH 7.5, containing 7 mol/L urea, 2% sarkosyl, and 0.35 mol/L NaCl. DNA was sheared by passage through a syringe needle (21 and 25 gauge), and the 32P-labeled nuclear RNA was isolated on a 5.7 mol/L CsCl gradient by ultracentrifugation at 100 000g for 18 hours at 20°C and then hybridized with cDNAs for PAI-1 and GAPDH (constitutive control) immobilized on nitrocellulose filters. The preparation of nitrocellulose filters containing the cDNAs, hybridization, and washing of the filters were carried out as described.34 The radioactivity corresponding to each individual filter was quantified by phosphorimaging autoradiography using a Molecular Dynamics Series 425F PhosphorImager in combination with ImageQuant software (Molecular Dynamics).
Fibrinolytic (Plasmin) Activity Assay
Surface-localized fibrinolytic activity was measured by using live, confluent, cultured HUVECs by direct conversion of HUVEC-bound, single-chain, 125I-labeled Glu-Pmg (by receptor-bound Pmg activators) to two-chain 125I-labeled plasmin and subsequent quantification of either 125I-labeled plasmin, Mr 20-kD light-chain, or Mr 60-kD heavy-chain formation, after SDS-PAGE under reducing conditions according to Mussoni et al35 as modified in this laboratory.1 19 Postconfluent, cultured HUVECs in 96-well plates (in triplicate) were incubated at 37°C for 24 hours in the absence or presence of HTG-VLDL (20 μg/mL) or Lp(a) (50 μg/mL) in serum-free culture medium. Saturating levels of TPA (200 nmol/L) were then added directly to the medium of each treated culture and incubated at 37°C for an additional 2 hours to titrate (form complexes) the newly produced PAI-1 induced by the lipoproteins. The cultures were then washed three times with 0.01 mol/L HEPES and 0.1 mol/L sodium acetate, pH 7.4, containing 1% BSA (buffer A) to remove TPA–PAI-1 complexes and unbound TPA and equilibrated with buffer A (50 μL per well) at 4°C for 20 minutes. 125I-labeled Glu-Pmg (2 μmol/L) in buffer A containing 1000 KIU/mL aprotinin (40 μL) was added to each well and incubated at 4°C for 30 minutes. Culture plates were then placed in a 37°C water bath to initiate the residual receptor-bound, TPA-mediated conversion of HUVEC-bound 125I-labeled Glu-Pmg to 125I-labeled plasmin. After a 10-minute incubation the reactions were stopped by the rapid addition of 40 μL of hot (56°C) solubilizing buffer (4% SDS, 10% glycerol, 0.2 mol/L Tris-HCl, pH 6.8). The total contents of each well were analyzed by 0.1% SDS-PAGE under reducing conditions.19 The amounts of 125I-labeled Mr 20-kD light-chain or Mr 60-kD heavy-chain plasmin generated were quantified by measuring the radioactivity content in each band by phosphorimaging autoradiography using a Molecular Dynamics Series 425F PhosphorImager in combination with ImageQuant software (Molecular Dynamics) as described previously.1 19
Analysis of Data
All of the data were expressed as mean±SD of replicate experiments performed in each assay and analyzed by Student’s t test. Data with P<.05 were taken to represent statistically significant differences in experimental results.
Determination of PAI-1 Genotypes (HindIII RFLPs) in Individual Human Umbilical Cords
Southern blot analysis of genomic DNA from different individual human umbilical cords, after digestion with HindIII and probing with a 32P-labeled 2.2-kb PAI-1 cDNA, identified the presence of the three distinct HindIII RFLPs designated 1/1 (22-kb band only), 1/2 (18- plus 22-kb bands), and 2/2 (18-kb band only), as described previously by Klinger et al31 (Fig 1⇓). The effects of HTG-VLDL or Lp(a) on the expression of PAI-1 antigen and mRNA were subsequently examined and compared in the individual PAI-1 genotyped HUVEC cultures and derived from their respective individual PAI-1 genotyped cords to determine specifically whether HTG-VLDL or Lp(a) induced similar or different responses in PAI-1 expression in these different PAI-1 genotyped HUVEC cultures.
Induction of PAI-1 Antigen and mRNA Expression by HTG-VLDL or Lp(a) in PAI-1 Genotyped Cultured HUVECs
Confluent, cultured HUVECs representing the three PAI-1 genotypes were analyzed separately for their individual expression of PAI-1 antigen and coincident mRNA levels in the absence or presence of HTG-VLDL (0 to 50 μg/mL) or Lp(a) (0 to 50 μg/mL). Both PAI-1 antigen and mRNA levels showed a time- and dose-dependent increase in HUVEC cultures with each of the lipoproteins, primarily in the 2/2 PAI-1 genotype (data not shown), and reached maximum levels of expression with HTG-VLDL and Lp(a) at 20 and 50 μg/mL, respectively. After incubation for 24 hours in the absence or presence of these lipoproteins, secreted PAI-1 antigen levels were increased ≈2- to 3-fold or ≈1.5- to 2-fold, respectively, in the 2/2 but not in the 1/2 or 1/1 PAI-1 genotype HUVEC cultures compared with each individual culture’s respective medium-incubated controls (Fig 2⇓). The basal levels of PAI-1 antigen secreted in 24 hours by each of the three PAI-1 genotype control HUVEC cultures were not significantly different (1360±150 for 1/1, 1210±170 for 1/2, and 1120±140 for 2/2, all in nanograms per milliliter per well). Simultaneous RPA analysis of these genotyped HUVEC cultures for their coincident expression of relative PAI-1 mRNA levels (PAI-1 mRNA to GAPDH mRNA ratios) indicated that PAI-1 mRNA levels reached their maximum at ≈8 to 10 hours in these HUVEC cultures (data not shown). After incubation for 8 hours in the absence or presence of HTG-VLDL or Lp(a), PAI-1 mRNA levels increased ≈3- to 4.5-fold or ≈2.5- to 3.2-fold, respectively, in the 2/2 but not in the 1/2 or 2/2 PAI-1 genotype HUVEC cultures (Fig 3⇓) compared with their respective medium-incubated controls. These combined PAI-1 antigen and mRNA results indicated that the 2/2 PAI-1 genotype was the most responsive to both HTG-VLDL and Lp(a) and that these lipoproteins appeared to induce PAI-1 expression in a genotype-specific manner. Simultaneous analysis of another fibrinolytic protein, UPA, indicated that HTG-VLDL and Lp(a) had no effect on either UPA antigen or mRNA levels compared with individual medium-incubated controls, suggesting that the effects of these lipoproteins on 2/2 PAI-1 expression do not represent a generalized stimulation phenomenon in cultured HUVECs (data not shown). These studies were carried out in eight and seven separate experiments for antigen and mRNA analysis, respectively, in which three or four different individual HUVEC cultures representing each of the three PAI-1 genotypes were examined and compared and yielded similar results.
To determine whether the genotype-specific induction of PAI-1 mRNA expression in 2/2 PAI-1 genotype cultured HUVECs, in response to HTG-VLDL or Lp(a), represented de novo RNA synthesis and/or required new protein synthesis, the effects of actinomycin D and puromycin on PAI-1 mRNA expression in cultured HUVEC was examined. Cultured HUVECs representing the 2/2 and 1/1 PAI-1 genotypes were preincubated with inhibitors for 30 minutes after an additional 8-hour incubation in the absence or presence of HTG-VLDL or Lp(a), and PAI-1 mRNA levels were measured by RPA. Actinomycin D and puromycin together completely abolished the increase in PAI-1 mRNA expression induced by either HTG-VLDL or Lp(a), whereas each inhibitor alone decreased PAI-1 mRNA expression only minimally (Fig 4⇓). These combined-inhibition results suggest that HTG-VLDL or Lp(a) induces the concomitant synthesis of other proteins that are necessary for the subsequent induction of PAI-1 expression. Neither actinomycin D nor puromycin had any inhibitory effect on the expression of GAPDH mRNA in these experiments. These inhibition studies were carried out in three separate experiments using three different individual 1/1, 1/2, and 2/2 genotype HUVEC cultures (in duplicate) and yielded similar results.
Effect of HTG-VLDL or Lp(a) on PAI-1 Gene Transcription Rates
To further establish whether the observed increase in PAI-1 mRNA in response to HTG-VLDL or Lp(a) was due to an increase in the rate of PAI-1 transcription rather than stabilization of the PAI-1 mRNA, nuclear transcription run-on assays were carried out. Nuclei were isolated from postconfluent 1/1 and 2/2 PAI-1 genotype cultured HUVECs (in 7500-mm2 flasks) and incubated in the absence or presence of HTG-VLDL, Lp(a), or IL-1 (positive control) for 8 hours at 37°C. Newly synthesized 32P-labeled nuclear transcripts were hybridized to PAI-1 and GAPDH cDNAs immobilized on nylon membranes, and the nuclear run-on assay results were measured by phosphorimaging autoradiography. The results indicated that transcriptional activation of the 2/2 PAI-1 genotype, as evidenced by a significant ≈5- to 6-fold or an ≈4- to 5-fold rise in de novo 32P-labeled PAI-1 mRNA (PAI-1 mRNA to GAPDH mRNA ratio) in the 2/2 PAI-1 genotype cultured HUVECs stimulated by HTG-VLDL or Lp(a), respectively, increased when compared with 1/1 PAI-1 genotype HUVEC cultures or medium-incubated controls. Results from a typical nuclear run-on assay are shown in Figs 5⇓ and 6⇓. The positive control IL-1 increased both 2/2 and 1/1 PAI-1 mRNA levels ≈5- to 6-fold. Nuclear transcription run-on results indicated the genotype-specific transcriptional regulation of the 2/2 PAI-1 genotype by both HTG-VLDL and Lp(a). These run-on assays were carried out in seven separate experiments using five different individual 1/1 and 2/2 PAI-1 genotype HUVEC cultures (in triplicate) and yielded similar results.
Surface-Localized Fibrinolytic Activity Assay
These experiments were carried out to determine whether the genotype-specific increase in PAI-1 levels induced by HTG-VLDL or Lp(a) exerted additional inhibitory effects on the net expression of cultured HUVEC surface-localized fibrinolytic activity. PAI-1 can regulate cell surface-localized fibrinolytic activity through its ability to complex and inactivate endogenous cultured HUVEC-produced and -bound PAs. To assess the potential contribution of increased PAI-1 activity levels induced by lipoproteins versus controls on cultured HUVEC surface-localized fibrinolytic activity, newly induced (24-hour) PAI-1 activity was titrated by addition of exogenous TPA, and the residual, uncomplexed, HUVEC-bound TPA activity was then measured by activation of cultured, HUVEC-bound 125I-labeled Glu-Pmg. HTG-VLDL and Lp(a) decreased cultured HUVEC-bound, TPA-mediated fibrinolytic activity by ≈50% to 60% in 2/2 PAI-1 but not in 1/2 or 1/1 PAI-1 genotype cultured HUVECs when compared with their respective medium-incubated controls (Fig 7⇓). The significant decrease in 2/2 PAI-1 genotype cultured HUVEC surface-localized fibrinolytic activity, as measured in this TPA titration assay, presumably reflects the net consequence of increased PAI-1 expression and subsequent interaction with endogenous surface-localized HUVEC TPA. These activity assays were carried out in five separate experiments using four different individual 1/1, 1/2, and 2/2 PAI-1 genotype HUVEC cultures (in duplicate) and yielded similar results.
An interrelationship between PAI-1 levels, PAI-1 polymorphisms (genotypes), and regulation by triglycerides was suggested by Dawson et al.16 The purpose of our present studies was to examine this hypothesized interrelationship in a PAI-1 genotyped (HindIII RFLP) cultured HUVEC model system. In these studies, we have now confirmed this interrelationship and further demonstrated that both HTG-VLDL and Lp(a) will induce PAI-1 antigen and mRNA expression in a genotype-specific manner, with the 2/2 PAI-1 genotype being the most responsive to these risk factor–associated components. In addition, combined results from inhibitor experiments and nuclear transcription run-on assays further indicated that these lipoproteins affected 2/2 PAI-1 gene expression at the level of transcription.
Impaired fibrinolytic activity may play an important role in the pathophysiological mechanisms underlying the initiation of atherosclerosis as well as the thrombotic complications associated with this disease, including MI.3 4 5 HTG has been associated with impaired fibrinolytic activity, which may be attributed in part to the coexistent elevation in plasma PAI-1 levels.4 16 36 37 38 Elevated levels of PAI-1 may therefore contribute to the increased risk for the thrombotic complications, CAD, and MI that are often observed in HTG individuals. Several in vitro studies have demonstrated that HTG-VLDL and Lp(a) will induce PAI-1 production. HTG-VLDL but not normal VLDL will induce PAI-1 antigen and mRNA expression in cultured HUVECs (1.6- to 2.5-fold).39 Lp(a) has been shown to increase PAI-1 antigen, activity, and steady-state mRNA levels without altering TPA activity or mRNA transcript levels in cultured HUVECs.40 We have confirmed these observations in cultured HUVECs and additionally demonstrated that these risk factor–associated components regulate PAI-1 gene expression in a genotype-specific manner as discussed below.
Several epidemiological studies have described an interrelationship between plasma PAI-1 levels, specific PAI-1 genotypes, and risk factors.14 16 41 The existence of polymorphic variations (genotypes) at three different sites in the human PAI-1 gene have been reported: a HindIII RFLP due to a base change at the 3′UTR,16 an allelic variation in the (CA)n dinucleotide repeat in the fourth intron,16 and a 4G/5G insertion/deletion polymorphism in the promoter region.13 The HindIII RFLP and (CA)n dinucleotide repeat region have been correlated with plasma PAI-1 and triglyceride levels, suggesting that triglyceride regulation of PAI-1 is genotype specific.16 Subjects homozygous for the 4G allele were found to have significantly higher PAI-1 activity than heterozygous (4G/5G) or homozygous 5G subjects.41 42 A recent study (ECTIM),15 however, suggested that the 4G allele was not a genetic risk factor for MI but might influence an individual’s fibrinolytic capacity and thus contribute to the risk profile, perhaps by interacting with other genetic and environmental factors. Our in vitro data have now further confirmed the hypothesized in vivo interrelationships between PAI-1, PAI-1 genotypes, and triglycerides through the use of a PAI-1 genotyped cultured HUVEC model and have clearly demonstrated the genotype-specific regulation of PAI-1 expression by risk factor associated–components, HTG-VLDL and Lp(a), with the 2/2 PAI-1 genotype (HindIII RFLP) being the most responsive to these lipoprotein inducers.
Several studies have demonstrated that regulation of the PAI-1 gene by a wide variety of inducer molecules can occur at the transcription level through specific cis-regulatory regions, including transforming growth factor-β,43 44 phorbol myristate acetate,45 and glucocorticoids.46 The exact mechanism by which HTG-VLDL and/or Lp(a) regulates PAI-1 expression in a genotype-specific manner has not been clearly identified or defined and will have to be examined further. Inhibitor (actinomycin D and puromycin) experiments and nuclear transcription run-on assays have, however, suggested that these atherogenic lipoproteins regulate 2/2 PAI-1 gene expression in cultured HUVECs at the transcriptional level rather than by stabilization of the PAI-1 mRNA. These results are clearly different from those observed in HepG2 cells, in which VLDL increased PAI-1 levels by stabilizing the steady-state level of PAI-1 mRNA rather than by increasing gene transcription.47 The significance of this difference in regulation of PAI-1 expression by VLDL in ECs versus HepG2 cells is unclear at present. Various other cell types have been shown to produce PAI-1, including smooth muscle cells,48 osteoblastic cells,49 lung fibroblasts,50 lung and kidney epithelial cells,51 and various tumor cell types.52 53 It is conceivable that the regulatory mechanisms affecting PAI-1 expression in different cell types may vary and may be closely associated with cellular function, thus requiring considerable site and cellular specificity.
Plasma levels of PAI-1 represent the total cumulative expression of PAI-1 from different cell types in response to a variety of inducers or a combination of inducers, including various combinations of CAD risk factors or risk factor–associated components. In view of these complex multicomponent interactions, it may be extremely difficult to dissect specific risk factor–associated effects or site-specific disease relationships. However, the studies described herein do provide strong support for the concept that individuals with the responsive 2/2 PAI-1 genotype may have the inherent (genetic) predisposition for HTG-VLDL- and/or Lp(a)-induced fibrinolytic dysfunction, promoting thrombotic complications, CAD, and atherothrombosis. In recent studies we have also demonstrated a strong association between the extent of angiographically identified CAD (two- or three-vessel disease) in symptomatic patients, homozygous PAI-1 genotypes (HindIII RFLPs, in particular the 2/2 PAI-1 genotype), and triglycerides.54 Ongoing studies in this laboratory will continue to further examine whether PAI-1 genotypes may be useful as noninvasive screening tools to predict increased thrombotic risk associated with HTG and elevated Lp(a) levels, as well as the potential for extensive CAD in certain symptomatic patients with the “high-risk” 2/2 PAI-1 genotype, and allow earlier initiation of secondary preventative measures.
Selected Abbreviations and Acronyms
|CAD||=||coronary artery disease|
|(HUV)ECs||=||(human umbilical vein) endothelial cell|
|PAI-1||=||plasminogen activator inhibitor 1|
|RFLP||=||restriction fragment length polymorphism|
|RPA||=||ribonuclease protection assay|
|TPA||=||tissue-type plasminogen activator|
|UPA||=||urokinase-type plasminogen activator|
This work was supported in part by National Institutes of Health grants HL-52791 (to F.M.B.), HL-44480 (to S.H.G.), and RR-00032 (to S.H.G.) (Bethesda, Md). Drs Li and Benza were supported in part by Institutional National Research Service award T32 HL-07703 during the course of this study. We would like to thank the labor and delivery staff at AMI Brookwood Medical Center and the University Hospital in Birmingham, Ala, for providing the umbilical cords used for isolation of HUVECs in these studies. We thank the staff of the University of Alabama at Birmingham General Clinical Research Center for drawing blood from our volunteer donors for isolation of HTG-VLDLs.
Li XN, Varma VK, Parks JM, Benza RL, Koons JC, Grammer JR, Grenett HE, Tabengwa EM, Booyse FM. Thrombin decreases the urokinase receptor and surface-localized fibrinolysis in cultured endothelial cells. Arterioscler Thromb Vasc Biol. 1995;15:410-419.
Haddock RC, Baker CD, Grammer JR, Parks JM, Speidel MT, Spell ML, Booyse FM. Urokinase binding and receptor identification in cultured endothelial cells. J Biol Chem. 1991;266:21466-21473.
Eliasson B, Attvall S, Taskinen M, Smith U. The insulin resistance syndrome in smokers is related to smoking habits. Arterioscler Thromb. 1994;14:1946-1950.
Auwerx JH, Bouillon R, Collen D, Geboers J. Tissue-type plasminogen activator antigen and plasminogen activator inhibitor in diabetes mellitus. Arteriosclerosis. 1988;8:68-72.
Bachmann FW. Fibrinolysis. In: Verstraete M, Vermylen J, Lijnen HR, Arnout J, eds. Thrombosis Haemostasis. Leuven, Belgium: Leuven University Press; 1987:227-265.
Sprengers ED, Kluft C. Plasminogen activator inhibitors. Blood. 1987;69:381-387.
Gray RP, Yudkin JS, Patterson DL. Plasminogen activator inhibitor: a risk factor for myocardial infarcion in diabetic patients. Br Heart J. 1993;69:228-232.
Dawson SJ, Wiman B, Hamsten A, Green F, Humphries S, Henney AM. The two allele sequences of a common polymorphism in the promoter of the plasminogen activator inhibitor-1 (PAI-1) gene respond differently to interleukin-1 in HepG2 cells. J Biol Chem. 1993;268:10739-10745.
Mansfield MW, Stickland MH, Grant PJ. Plasminogen activator inhibitor-1 (PAI-1) promoter polymorphism and coronary artery disease in non-insulin-dependent diabetes. Thromb Haemost. 1995;74:1021-1034.
Ye S, Green FR, Scarabin PY, Nicaud V, Bara L, Dawson SJ, Humphries SE, Evans A, Luc G, Cambou JP, Arveiler D, Henney AM, Cambien F. The 4G/5G genetic polymorphism in the promoter of the plasminogen activator inhibitor-1 (PAI-1) gene is associated with differences in plasma PAI-1 activity but not with risk of myocardial infarction in the ECTIM Study. Thromb Haemost. 1995;74:837-841.
Dawson SJ, Hamsten A, Wiman B, Henney A, Humphries S. Genetic variation at the plasminogen activator inhibitor-1 locus is associated with altered levels of plasma plasminogen activator inhibitor-1 activity. Arterioscler Thromb. 1991;11:183-190.
Maciag T, Cerundolo J, Ilsley S, Kelley PR, Forand R. An endothelial cell growth factor from bovine hypothalamus: identification and partial characterization. Proc Natl Acad Sci U S A. 1979;76:5674-5678.
Voyta JC, Via DP, Butterfield CE, Zetter BR. Identification and isolation of endothelial cells based on their increased uptake of acetylated-low density lipoprotein. J Cell Biol. 1984;99:2034-2040.
Booyse FM, Osikowicz G, Feder S, Scheinbuks J. Isolation and characterization of a urokinase-type plasminogen activator (Mr=54,000) from cultured human endothelial cells indistinguishable from urinary urokinase. J Biol Chem. 1984;259:7198-7205.
Booyse FM, Scheinbuks J, Lin P, Traylor M, Bruce R. Isolation and interrelationships of the multiple molecular tissue-type and urokinase-type plasminogen activator forms produced by cultured human umbilical vein endothelial cells. J Biol Chem. 1988;263:15129-15138.
Jaffe EA, Nachman RL, Becker CG, Minick CR. Culture of human endothelial cells derived from umbilical veins: identification by morphologic and immunologic criteria. J Clin Invest. 1973;52:2745-2756.
Thornton SC, Mueller SN, Levine EM. Human endothelial cells: use of heparin in cloning and long-term serial cultivation. Science. 1983;222:623-625.
Booyse FM, Quarfoot AJ, Chediak J, Stemerman MB, Maciag T. Characterization and properties of cultured human von Willebrand umbilical vein endothelial cells. Blood. 1981;58:788-796.
Booyse FM, Quarfoot AJ, Bell S, Fass DN, Lewis JC, Mann KG, Bowie DJW. Cultured aortic endothelial cells from pigs with von Willebrand’s disease:in vitro model for studying the molecular defect(s) of the disease. Proc Natl Acad Sci U S A. 1977;74:5702-5706.
Booyse FM, Sedlak BJ, Rafelson ME Jr. Culture of arterial endothelial cells: characterization and growth of bovine aortic cells. Thromb Diasthes Haemorrh. 1975;34:825-839.
Maniatis T, Fritsch EF, Sambrook J. Synthetic oligonucleotide probes. In: Irwin N, Ford N, Nolan C, Ferguson M, Ockler M, eds. Molecular Cloning: A Laboratory Manual, 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory; 1982:9.16-9.23.
Klinger KW, Winqvist R, Riccio A, Andreasen PA, Sartorio R, Nielsen LS, Stuart N, Stanislovitis P, Watkins P, Douglas R, Grzeschik K, Alitalo K, Blasi F, Dano K. Plasminogen activator inhibitor type 1 gene is located at region q21.3-q22 of chromosome 7 and genetically linked with cystic fibrosis. Proc Natl Acad Sci U S A. 1987;84:8548-8552.
Ny T, Sawdey M, Lawrence D, Millan JL, Loskutoff DJ. Cloning and sequence of a cDNA coding for the human B-migrating endothelial-cell-type plasminogen activator inhibitor. Proc Natl Acad Sci U S A. 1986;83:6776-6780.
Fless GM, Snyder ML. Polymorphic forms of Lp(a) with different structural and functional properties: cold induced self-association, and binding to fibrin and lysine-Sepharose. Chem Phys Lipids. 1994;67/68:69-79.
Tso JY, Sun X, Kao T, Reece KS, Wu R. Isolation and characterization of rat and human glyceraldehyde-3-phosphate dehydrogenase cDNAs: genomic complexity and molecular evolution of the gene. Nucleic Acids Res. 1985;13:2485-2502.
Paramo JA, Colucci M, Collen D. Plasminogen activator inhibitor in the blood of patients with coronary artery disease. Br Med J. 1985;291:573-574.
Stiko-Rahm A, Wiman B, Hamsten A, Nilsson J. Secretion of plasminogen activator inhibitor-1 from cultured human umbilical vein endothelial cells is induced by very low density lipoproteins. Arteriosclerosis. 1990;10:1067-1073.
Etingin OR, Hajjar DP, Hajjar KA, Harpel PC, Nachman RL. Lipoprotein (a) regulates plasminogen activator inhibitor-1 expression in endothelial cells: a potential mechanism in thrombogenesis. J Biol Chem. 1991;268:2459-2465.
Eriksson P, Kallin B, van “T Hooft FM, Bavenholm P, Hasten A. Allele-specific increase in basal transcription of the plasminogen-activator inhibitor 1 gene is associated with myocardial infarction. Proc Natl Acad Sci U S A. 1995;92:1851-1855.
Panahloo A, Mohamed-Ali V, Lane A, Green F, Humphries SE, Yudkin JS. Determinants of plasminogen activator inhibitor 1 activity in treated NIDDM and its relations to a polymorphism in the plasminogen activator inhibitor 1 gene. Diabetes. 1995;44:37-42.
Keeton MR, Curriden SA, van Zonneveld AJ, Loskutoff DJ. Identification of regulatory sequences in the type I plasminogen activator inhibitor gene repsonsive to transforming growth factor β. J Biol Chem. 1991;266:23048-23052.
Westerhausen DR, Hopkins WE, Billadello JJ. Multiple transforming growth factor-β-inducible elements regulate expression of the plasminogen activator inhibitor type-1 gene in Hep G2 cells. J Biol Chem. 1991;266:1092-1100.
Descheemaeker KA, Wyns S, Nelles L, Auwerx J, Ny T, Collen D. Interaction of AP-1-, AP-2-, and Sp1-like proteins with two distinct sites in the upstream regulatory region of the plasminogen activator inhibitor-1 gene mediates the phorbol 12-myristate 13-acetate response. J Biol Chem. 1992;267:15086-15091.
van Zonneveld AJ, Curriden SA, Loskutoff DJ. Type 1 plasminogen activator inhibitor gene: functional analysis and glucocorticoid regulation of its promoter. Proc Natl Acad Sci U S A. 1988;85:5525-5529.
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.
Thalacker FW, Nilsen-Hamilton M. Opposite and independent actions of cyclic AMP and transforming growth factor beta in the regulation of type 1 plasminogen activator inhibitor expression. Biochem J. 1992;287:855-862.
de Vries TJ, Kitson JL, Silvers WK, Mintz B. Expression of plasminogen activators and plasminogen activator inhibitors in cutaneous melanomas of transgenic melanoma-susceptible mice. Cancer Res. 1995;55:4681-4687.
Benza RL, Grenett HE, Li XN, Grammer JR, Tabengwa EM, Brown SL, Perry GJ, Holman WL, McGiffin DC, Booyse FM. Association of fibrinolytic protein gene polymorphisms with certain risk factors for coronary artery disease: genetic markers for increased thrombotic risk? Circulation. 1995;92(suppl I):I-494. Abstract.