Five Frequent Polymorphisms of the PAI-1 Gene
Lack of Association Between Genotypes, PAI Activity, and Triglyceride Levels in a Healthy Population
Abstract The main function of plasminogen activator inhibitor type 1 (PAI-1) is to decrease fibrinolysis, which leads to fibrin accumulation. An elevated plasma PAI-1 concentration has been identified as a risk factor for the development of myocardial infarction, and an association between 1 polymorphism of the PAI-1 promoter and plasma PAI-1 levels has been described. Our aim was to identify new polymorphisms in the PAI-1 gene and to further examine the relationship between PAI-1 genotypes and circulating PAI-1 levels. We report the presence of 4 new polymorphisms that were identified by nonisotopic single-strand conformational polymorphism analysis followed by sequencing. These polymorphisms were investigated in relation to PAI-1 levels in a sample of 256 healthy men, aged 50-59 years, from France and Northern Ireland. Two G/A substitutions were detected at positions −844 and +9785. The former is in strong positive linkage disequilibrium with the previously described 4G/5G polymorphism at position −675. Two polymorphisms in the 3′ untranslated region were identified. One corresponds to a T/G substitution at position +11 053 and is in negative linkage disequilibrium with the G/A substitution (+9785). The other is a 9-nucleotide insertion/deletion located between nucleotides +11 320 and +11 345 in a threefold-repeated sequence. This polymorphism is in strong positive linkage disequilibrium with the G/A substitution (+9785). The overall heterozygosity provided by the 5 PAI-1 polymorphisms (including the 4 new variants and the 4G/5G polymorphism) was .77. No significant association was found between PAI activity and genotypes; furthermore, the well known associations between PAI activity and body mass index, serum triglycerides, or insulin were homogeneous according to PAI-1 genotypes.
- Received June 11, 1996.
- Accepted August 27, 1996.
Plasminogen activator inhibitor-1 (PAI-1) is a 50-kD glycoprotein that belongs to the serine protease inhibitor superfamily (serpins)1 2 3 and is the primary inhibitor of both tissue-type and urokinase-type PAs.4 5 The literature argues that PAI-1 is implicated in coronary heart disease (reviewed in Reference 66 ). Reduced plasma fibrinolytic activity, mainly due to elevated PAI activity, has been observed in patients with angina pectoris or in those with a history of MI. In one study, a high plasma PAI activity independently predicted reinfarction within 3 years of the primary event.7 The plasma concentration of PAI-1 antigen was found to be a reliable predictor of coronary events in patients with angina pectoris.8 It was also recently shown that PAI levels were related to the extent of vessel wall atherosclerosis.9
Until now, impaired fibrinolytic activity due to increased PAI-1 levels was thought to represent an acquired abnormality, with intraindividual variability dependent on the individual metabolic or inflammatory status.10 Indeed, many cross-sectional studies have shown that plasma PAI-1 levels were positively correlated with BMI, WHR, fasting insulinemia, VLDL-TG concentrations and, to a lesser extent, fibrinogen.
Gene variability could also contribute to the variability in circulating PAI-1 levels. Four different polymorphisms on the PAI-1 gene have been described: two (CA)n repeat polymorphisms, one in the promoter and one in intron 411 12 ; an HindIII RFLP13 ; and an insertion (5G)/deletion (4G) polymorphism at position −675 of the PAI-1 promoter.14 Significant associations between polymorphisms of the PAI-1 gene and plasma PAI-1 levels have been reported. In MI patients and control subjects, higher plasma PAI-1 levels were significantly associated with the smaller alleles of the intron 4 (CA)n repeat. Studies of the HindIII RFLP, which is in strong linkage disequilibrium with the (CA)n repeat, demonstrated that plasma PAI-1 levels were lower in both control subjects and MI patients with an additional HindIII site.12 Studies of the 5G/4G polymorphism have shown a higher plasma PAI activity in subjects with the 4G than with the 5G allele in MI patients,14 15 16 non–insulin-dependent diabetics,17 and healthy control subjects.15 16 In addition, a stronger association between plasma fibrinogen14 or TG levels17 18 and plasma PAI activity has been observed in individuals homozygous for the 4G allele. In a recent study in Sweden,15 the prevalence of the 4G allele was significantly higher in a group of 100 MI patients aged 35 to 45 years than in age-matched control subjects. However, this relation was not confirmed in a large study of MI patients aged 25 to 64 years.16 Also, no difference was observed in the frequency of HindIII alleles or intron 4 (CA)n repeats between control and type 1 or type 2 diabetic subjects with or without retinopathy.11 We now report the results of an extensive analysis of polymorphisms of the PAI-1 gene in relation to plasma PAI activity levels and metabolic status in a population of healthy men.
Subjects were participants in a cohort study of risk factors for MI (PRIME Study), the details of which are described elsewhere (the PRIME Study Group, unpublished data). Men aged 50 to 59 years were recruited from both industrial and community settings in three cities of France and Belfast, Northern Ireland. A random sample of 256 healthy subjects was selected from this cohort for the present study. Plasma PAI activity was investigated in 245 of them. The subjects provided informed consent to participate in the study.
BMI (weight in kilograms divided by the square of height in meters) was calculated. Fat distribution was assessed by using the WHR, which is the ratio of the waist circumference at the level of the umbilicus to that of the hips at the level of the greater trochanter and symphysis pubis.
Blood was collected between 8 and 10 AM (into citrate-containing tubes) to overcome the diurnal variation in PAI-1 levels and processed immediately. After 30 minutes of centrifugation (2500g) at 4°C, the middle layer of the plasma was rapidly pipetted off and stored at −80°C. PAI activity was assayed by a commercially available kit (Spectrolyse/PL Biopool). Plasma insulin was measured with a commercially available radioimmunoassay kit (CEA SORIN) and plasma TGs by the method of Buccolo and David.19
Search for Novel Polymorphisms by PCR-SSCP and Sequencing
Genomic DNA from 37 white volunteer donors recruited from the medical and laboratory staff was used to screen the PAI-1 gene for polymorphisms. Genomic DNA was extracted from peripheral blood leukocytes by the salting-out method.20 The nt sequence reported in this article has been published by Bosma et al21 and can be obtained from the GenBank/EMBL databank, access No. J03764 M55991.
The PAI-1 promoter was divided into five overlapping fragments (P1 to P5) that covered the gene sequence from nt −1503 to nt +160. The 9 exons and short portions of the flanking introns were divided into 9 parts (E1 to E9). In the 3′ untranslated region, the sequence from nt +10 977 to nt +12 322 was divided into 5 parts (3′1 to 3′5; Fig 1⇓). Primer sequences, lengths, locations of amplified fragments, and annealing temperatures are reported in Table 1⇓ for the promoter, Table 2⇓ for exons, and Table 3⇓ for the 3′ untranslated region. Primers were synthesized by Eurogentec. PCR was performed in a total volume of 50 μL containing 125 ng genomic DNA, 190 μmol/L dNTP, 5 μmol/L 16-dUTP/biotin (Boehringer Mannheim), 12.5 pmol/L of each primer, and 0.75 U Taq polymerase (Bioprobe Systems) in 1× Taq polymerase buffer [20 mmol/L Tris HCl, pH 8.5; 16 mmol/L (NH4)2SO4; 2.5 mmol/L MgCl2; and 150 μg/mL BSA]. Samples were processed through 40 cycles of 45 seconds at 97°C; 1 minute at each annealing temperature (established for each primer couple); 1 minute, 30 seconds at 72°C; and a terminal 72°C extension for 5 minutes on a GeneAmp PCR system 9600 (Perkin-Elmer). Amplified DNA fragments larger than optimal size for SSCP were digested with appropriate restriction endonucleases (Tables 1⇓, 2⇓, and 3⇓) according to the manufacturer's instructions.
For SSCP, 5 μL of amplified products was mixed with 3 μL sample buffer (95% deionized formamide, 20 mmol/L EDTA [pH 8.0], 0.05% bromophenol blue, and 0.05% xylene cyanole). The samples were denatured for 5 minutes at 95°C, rapidly chilled on ice, and loaded onto a nondenaturing 6% or 8% polyacrylamide gel (acrylamide/bisacrylamide, 19:1, wt/wt) containing 10% or 7.5% glycerol, respectively. For the 3′2 fragment polymorphism, genotype was determined by migration of the PCR products under denaturing conditions (6% polyacrylamide gel, 7 mol/L urea). The gel was prepared in 1× TBE buffer (89 mmol/L Tris, 89 mmol/L boric acid, 2 mmol/L EDTA, pH 8.0). Electrophoresis was carried out in 1× TBE buffer at 1200 V for ≈5 hours; the migration temperature was maintained at 23°C. After electrophoresis, the DNA was transferred overnight to a positive nylon membrane (Hybond N+, Amersham International plc) with capillarity transfer in 20× SSC buffer (300 mmol/L sodium citrate, 3 mol/L NaCl, pH 7.0). After cross-linking was performed (2 J/cm2 at 312 nm), the biotin-labeled DNA was revealed as follows. The membrane was saturated with PBS/5% nonfat dried milk for 15 minutes, washed in PBS/Triton X-100 (0.01%), and incubated for 1 hour at room temperature in avidin-phosphatase solution (France Biochem) diluted 1/5000 in PBS/0.5% nonfat dried milk. After the membrane was washed, the final reaction product was visualized with NBT-BCIP (Sigma) as the substrate for 2 hours at room temperature.
At least 4 genomic DNAs of each SSCP profile were selected and sequenced. PCR was performed under the same conditions as for SSCP analysis but without 16-dUTP/biotin. PCR products were purified by phenol-chloroform deproteinization and ethanol precipitation. Amplified and purified DNAs were sequenced by ESGS or by the Centre commun de biologie moléculaire CHU Timone (Marseille, France).
Xho I Restriction Analysis
The genotypes resulting from the G→A substitution at position −844 of the promoter fragment P2 were assessed by Xho I endonuclease digestion. Genomic DNAs were amplified by PCR with the same primer couple as for SSCP analysis but without 16-dUTP/biotin. PCR products were digested with Xho I restriction enzyme (Eurogentec) according to the manufacturer's instructions and electrophoresed on a 2% agarose gel stained with ethidium bromide.
Allele-Specific Oligonucleotide Hybridization
The genotypes for polymorphisms P3, E8, and 3′1 were analyzed by allele-specific oligonucleotide hybridization with biotinylated probes. PCR products (10 μL) were denatured in 200 μL of 0.5 mol/L NaOH and 1.5 mol/L NaCl, and 100 μL of the dilution was spotted onto Hybond N+ nylon membranes by use of a dot blot apparatus and then neutralized with 5× SSC buffer. The filters were then cross-linked by UV irradiation, prehybridized (in 5× SSC; 5× Denhardt's solution; 20 mmol/L EDTA, pH 8.0; and 0.1% SDS), and hybridized with a biotin-labeled probe diluted to 1 pmol/mL of prehybridization buffer at an optimal temperature (empirically determined for each probe to be near its melting temperature) for 1 hour. The membranes were washed twice in 0.5× SSC at room temperature for 5 minutes and then once in 1× SSC for 5 minutes at Tm-3. The filters were then immersed for 30 minutes in saturation buffer (1% blocking reagent [Boehringer Mannheim] in maleic buffer [100 mmol/L maleic acid and 150 mmol/L NaCl, pH 7.5]). Probe hybridization was revealed by a 30-minute incubation with avidin–alkaline phosphatase diluted 1/20 000 in saturation buffer followed by three washes of 15 minutes each in washing buffer (maleic buffer and Triton X-100 [0.01%]) in an NBT-BCIP system as described above.
Biotin-labeled probes (15-mer oligonucleotides) were synthesized and labeled by Eurogentec. Probe sequences recognizing the 4G/5G polymorphism in PAI-1 promoter region P3 were CACGTGGGGAGTCA and ACGTGGGGGAGTCA, respectively. The new polymorphism, corresponding to an nt substitution (G→A) in E8, was analyzed with probe G (CACTCCCGCCTGGGC) or probe A (CACTCCCACCTGGGC). An nt substitution (G→T) in 3′1 was analyzed with probe G (GAAGAAAGGTCAGAT) or probe T (GAAGAAATGTCAGAT). The reliability of genotyping by allele-specific oligonucleotide hybridization was confirmed by blotting onto each membrane two sequenced PCR products of each genotype and two irrelevant amplified products as positive and negative controls, respectively.
Statistical analysis was performed with SAS software (SAS Institute Inc). Allele frequencies were estimated by gene counting. Departure from Hardy-Weinberg equilibrium was systematically checked by the χ2 test with df=1. Pairwise linkage disequilibrium coefficients were estimated in the sample. Coefficients were reported as the ratios of the unstandardized coefficients to their minimal/maximal values (‖D′‖). The sign before the coefficients indicates whether the linkage disequilibrium is positive or negative.22 The coefficient of heterozygosity was estimated by using the Myriad program.23 The distribution of PAI activity values was positively skewed, so logarithmically transformed values were used. Arithmetic means of PAI activity levels are given in the tables. ANOVA was used to detect any association between genotype and PAI activity. Mean values of PAI activity were compared in univariate analysis and after adjustment for relevant covariates. The homogeneity of results was tested before data from the French centers were pooled. ANCOVA was used to test for any interaction between genotype and determinants of PAI activity (ie, TGs, BMI, WHR, and insulin). Two-tailed values of P<.05 were considered significant.
Systematic Search for Polymorphisms of the PAI-1 Gene
Analysis of the Promoter: Identification of 1 nt Substitution at Position −844
The five overlapping regions from the first nt of the P1 fragment (nt −1503) to the last nt of the P5 fragment (nt +160), which largely overlapped the E1 fragment (Fig 1⇑), were amplified by PCR and subjected to nondenaturing 6% and 8% polyacrylamide gel electrophoresis. In the 37 subjects tested, no variation in electrophoretic mobility was found for fragments P1, P4, or P5.
The amplified products of the P2 fragment (from nt −1212 to nt −703) were digested with Taq I or Ava II enzyme before nondenaturing gel electrophoresis. Two fragments were obtained after digestion with Taq I (145 and 365 bp, Fig 2A⇓), and three smaller fragments (189, 242, and 79 bp) were obtained after digestion with Ava II (data not shown). Variations in electrophoretic mobility were seen in the 145- and 189-bp fragments. Direct sequencing of each polymorphic fragment identified a G→A transition at position −844. This substitution abolished a restriction site for the endonuclease Xho I. After digestion with this enzyme, the G/G genotype, characterized by the presence of two fragments of 314 and 146 bp, could be visualized on a 2% agarose ethidium bromide–stained gel, whereas genotype A/A was characterized by a single band of 510 bp. All three bands were present in heterozygous subjects (Fig 2B⇓).
Analysis of the Coding Region: Identification of 1 nt Substitution in the Intron Sequence Flanking Exon 8 at Position +9785
The entire coding region and a large portion of the flanking intron sequences from exons 1 to 9, which represents 2941 bp, were systematically screened by the PCR-SSCP technique as described in Fig 1⇑. Because intron 6 is short, exons 6 and 7 were analyzed in a single amplified fragment. No variations were observed in the electrophoretic profiles of single-strand and renatured double-strand amplified DNA on either a 6% or an 8% polyacrylamide gel with PCR products E1, E2, E3, E4, E5, E6-7, and E9, which correspond to the coding part of exon 9.
Amplified products of E8 (445 bp, from nt +9621 to nt +10 065) were digested with Pst I, yielding two fragments (135 and 310 bp). A shift in electrophoresis migration of single-strand and renatured double-strand DNA of the 310-bp fragment was observed (Fig 2D⇑). No modification in electrophoretic profile could be observed for the 135-bp single- and double-strand DNA fragments. Sequencing revealed a G→A substitution in nt +9785. In all sequenced DNA, a G insertion was found after nt +9789, which is inconsistent with the published sequence.21
Analysis of the 3′ Untranslated Region: Identification of Two Polymorphisms
The 3′ untranslated region of exon 9 was studied from nt +10 977 to nt +12 322. Owing to sequence complexity, the region from nt +11 633 to nt +12 029, which contains the sequence corresponding to the AU-rich 3′ untranslated region of mRNA (+11 791 to +11 876), was studied in both the 3′3 and 3′4 amplified fragments. No polymorphism was detected in 3′3, 3′4, and 3′5 amplified fragments (+11 997 to +12 322) that contain three potential polyadenylation signals for the 3.2-kb mRNA species24 (not shown).
However, the 3′1 PCR fragment (+10 977 to +11 259), which contains the potential polyadenylation signal (ATAAT, +11 163 to +11 167) for the 2.3-kb mRNA species, appeared to be polymorphic (Fig 2E⇑). Sequencing revealed a G→T substitution at position +11 053. Contrary to the published sequence, a T was always identified at positions +11 162 and +11 165. Dde I restriction analysis of the 37 tested DNAs confirmed the presence of T at position 11 165 (data not shown).
PCR-SSCP analysis of the amplified 3′2 fragment (349 bp) from nt +11 205 to +11 554 showed an important difference in the migration pattern in 7 of 37 analyzed samples (Fig 2F⇑). Sequencing revealed deletion of a sequence of 9 nt from a threefold-repeated sequence (CGCGCCCCC) between nt +11 320 and nt +11 345. Migration of the 3′2 fragment under denaturing conditions was used to distinguish heterozygotes from homozygotes (Fig 2G⇑). The different polymorphisms identified are reported in Table 4⇓.
Genotype Frequencies and Association Between Polymorphisms
Except for that of the 3′2 fragment, genotype distributions did not deviate from Hardy-Weinberg expectations. As shown in Table 5⇓, the polymorphisms in fragments P2 and P3 were in strong, positive linkage disequilibrium, as were those in E8 and 3′2 fragments. Conversely, a negative linkage disequilibrium or no association was observed between the other polymorphisms. The overall heterozygosity provided by the 5 PAI-1 polymorphisms was .77.
Association of Polymorphisms With Plasma PAI Activity Levels
Considering the entire population as well as the different geographic centers, we observed no relationship between genotype and plasma PAI activity, except for the E8 fragment polymorphism. A genotype effect ([AA+AG] versus GG) was observed in the French population (P<.05). In contrast, this relation was not found in the Irish population and, in fact, was exactly opposite (Table 6⇓). Adjustment for age, BMI, WHR, TGs, and insulin did not change these results. The relationship between PAI activity and serum TGs, BMI, WHR, and insulin was similar in the different PAI-1 genotypes. As an illustration, regression analysis of PAI activity on TG levels in the different PAI-1 genotypes is presented in Table 7⇓.
Gene variability is the source of interindividual differences and could explain the variation in susceptibility to atherosclerosis and MI. Characterization of the molecular variability of candidate genes could allow identification of populations at particular risk. PAI-1 has been implicated in the thrombotic phenomenon. Animal experiments have shown that modulation of PAI activity in plasma influences thrombus growth and thrombolysis.25 26 27 Recently, in vivo data have been accumulated about the role of PAI-1 in smooth muscle cell migration or proliferation secondary to a carotid lesion.28 PAI-1 is therefore an eminent candidate gene for MI development.
In this study, we identified 4 new polymorphisms in the PAI-1 gene by using the SSCP method. This technique29 can be assumed to detect >90% of 1-nt variations.30 We adapted and used nonisotopic conditions to obtain optimal sensitivity and reproducibility. However, we were confronted with an unresolved issue concerning the E2 fragment. When the SSCP technique revealed a shift in migration (data not shown), no nt variations were detected by sequencing.
The new polymorphism in the promoter region corresponds to a G→A substitution at position −844 and is in strong linkage disequilibrium with the previously described 4G/5G polymorphism at position −675. The latter affects the binding of nuclear proteins involved in the regulation of PAI-1 gene transcription. Although both alleles bind a transcriptional activator, the 5G allele also binds a repressor protein to an overlapping binding site. In the absence of the bound repressor, the basal level of PAI-1 gene transcription is increased.14 15 The substitution at position −844 is included in a consensus sequence binding site Ets nuclear protein. Ets proteins have been implicated in the regulation of gene expression in a variety of biological processes including growth control, transformation, T-cell activation, and development in many organisms (reviewed in Reference 3131 ). A DNA sequence similar to the one described in the PAI-1 gene promoter has already been described in genes that have been implicated in tissue remodeling.32 Interestingly, the Ets protein is highly expressed when angiogenesis resumes as well as during tumor vascularization. It would be interesting to establish whether the two observed alleles bind different nuclear proteins.
The second polymorphism described (a G→A substitution at position +9785) was located in intron 7 of the PAI-1 gene, 82 nt before the putative alternative splicing site 33 and 111 nt before the first nt coding for amino acids of the reactive center of PAI-1.34
Polymorphisms were also observed in the 3′ untranslated region. The first was found on fragment 3′1 and corresponds to a G→T substitution at position +11 053 (a T in this position has been previously described24 ). Sequencing of this region allowed us to identify a T at positions +11 162 and +11 165 according to the results of Ny et al2 and Ginsburg et al,1 which contrast with those of Bosma et al.21 This difference could constitute a rare polymorphism or a sequencing error.
The second polymorphism described in the 3′ untranslated region corresponds to a 9-nt deletion in a short sequence repeat between nt +11 319 and nt +11 345. These data are discordant with those of Dawson et al,14 who failed to identify a sequence change in this area by chemical cleavage mismatch analysis. This polymorphism and the one described in fragment E8 are in strong, positive linkage disequilibrium. Interestingly, the polymorphism present in fragment E8 was not in linkage disequilibrium with those in the promoter region but was strongly negatively linked to that on the 3′1 fragment. The allele frequency of polymorphism P3 was close to that located in the P2 fragment and was consistent with estimates previously published in healthy subjects.14 15 16
A higher plasma PAI activity level has been described in patients with the 4G/4G genotype. Although one group of non–insulin-dependent diabetics did not show an association between PAI-1 genotype and circulating level of PAI activity,18 others have observed such an association in patients with previous MI or non–insulin-dependent diabetes.14 15 16 17 In healthy subjects this relationship might be more elusive, since Dawson et al14 could not demonstrate it as being significant and Ye et al16 needed to pool different populations to reach a large number (n=601) of control subjects to obtain a significant difference. In our population sample of healthy subjects, no significant relation was observed between the various genetic polymorphisms and PAI activity except for the E8 polymorphism. Owing to the small number of AA genotypes analyzed, the genotype effect observed in the French population does not warrant a firm conclusion. These discrepant results could be due to the small size of our sample or the absence of a pathological trait such as diabetes. Furthermore, PAI-1 plasma levels may not reflect key changes in local tissue concentrations. Additional genetic differences in the study populations could also be implicated.
Significant interaction has been described between PAI-1 genotype and serum TG levels, particularly in non–insulin-dependent diabetic patients.17 18 Such a relation was not demonstrated in patients with a previous MI or healthy control subjects16 and was not found in our population. Perhaps the presence of non–insulin-dependent diabetes is a prerequisite for this interaction.
We conclude that the relationship between genotype and plasma PAI activity in healthy white subjects is weak. The discrepancy observed between the results in our study and others needs to be clarified. It might be attributable to a difference in clinical severity or geographic origin of the subjects analyzed.
Selected Abbreviations and Acronyms
|BMI||=||body mass index|
|PA(I)||=||plasminogen activator (inhibitor)|
|PCR||=||polymerase chain reaction|
|RFLP||=||restriction fragment length polymorphism|
|SSCP||=||single-strand conformational polymorphism|
This work was supported by grants from INSERM (CJF 93-12) and the Ministère du Travail et des Affaires Sociales, Direction des Hôpitaux, Programme Hospitalier de Recherche Clinique (PHRC 1994). The PRIME Study is supported by an agreement between INSERM and the Merck, Sharpe and Dohme–Chibret Laboratory with the following participating laboratories: the Bas-Rhin MONICA Project, Faculty of Medicine, Strasbourg (D. Arveiler, B. Haas); the Haute-Garonne MONICA Project, INSERM U326-ORS, Toulouse (J. Ferrières, J.B. Ruidavets); the Lille MONICA Project, Pasteur Institute, Lille (P. Amouyel, M. Montaye); the Belfast MONICA Project, The Queen's University of Belfast, Belfast, Northern Ireland (A. Evans, J. Yarnell); the SERLIA Laboratory, INSERM U325, and Pasteur Institute, Lille (G. Luc, J.M. Bard); the Laboratory of Haematology, La Timone Hospital, Marseilles (I. Juhan-Vague); the Laboratory of Hormonology, INSERM U326, Toulouse (B. Perret); the Trace Element Laboratory, Department of Medicine, The Queen's University of Belfast, Belfast, Northern Ireland (D. McMaster); the Vitamin Research Unit, University of Bern, Bern, Switzerland (F. Gey); the DNA Bank, Service Commun No. 7, INSERM, Paris (F. Cambien); and the Coordinating Center, INSERM U258, Paris (P. Ducimetière, P.Y. Scarabin, A. Bingham).
Ginsburg D, Zeheb R, Yang AY, Rafferty UM, Andreasen PA, Nielsen L, Dano K, Lebo R, Gelehrter TD. cDNA cloning of human plasminogen activator-inhibitor from endothelial cells. J Clin Invest. 1986;78:1673-1680.
Ny T, Sawdey M, Lawrence D, Millan JL, Loskutoff DJ. Cloning and sequence of a cDNA coding for the human β-migrating endothelial-cell-type plasminogen activator inhibitor. Proc Natl Acad Sci U S A. 1986;83:6776-6780.
Kruithof EKO, Tran-Thang C, Ransijn A, Bachmann F. Demonstration of the fast-acting inhibitor of plasminogen activators in human plasma. Blood. 1984;64:907-913.
Salomaa V, Stinson V, Kark JD, Folsom AR, Davis CE, Wu KK. Association of fibrinolytic parameters with early atherosclerosis: the ARIC study. Circulation. 1995;91:284-290.
Juhan-Vague I, Alessi MC, Joly P, Thirion X, Vague P, Declerck PJ, Serradimigni A, Collen D. Plasma plasminogen activator inhibitor-1 in angina pectoris: influence of insulin and acute phase response. Arteriosclerosis. 1989;9:362-367.
Dawson S, 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.
Klinger KW, Winqvist R, Riccio A, Andreasen PA, Sartorio R, Nielsen LS, Stuart N, Stanislovitis P, Watkins P, Douglas R, Grzeschik KH, Alitalo 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. 1991;84:8548-8552.
Dawson S, 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.
Eriksson P, Kallin B, van't Hooft FM, Bavenholm P, Hamsten 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.
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) 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.
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 relation to a polymorphism in the plasminogen activator inhibitor 1 gene. Diabetes. 1995;44:37-42.
Buccolo G, David M. Quantitative determination of serum triglycerides by use of enzymes. Clin Chem. 1973;19:475-482.
Miller SA, Dykes DD, Polesky HF. A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res. 1988;16:1215.
Bosma PJ, van den Berg EA, Kooistra T, Siemieniak DR, Slightom JL. Human plasminogen activator inhibitor-1 gene. J Biol Chem. 1988;263:9129-9141.
Nei M. Molecular Evolutionary Genetics. New York, NY: Columbia University Press; 1987.
Maclean CJ, Morton NE. Estimation of Myriad Haplotype frequencies. Gen Epidemiol. 1985;2:263-272.
Biemond BJ, Levi M, Coronel R, Janse MJ, Ten Cate JW, Pannekoek H. Thrombolysis and reocclusion in experimental jugular vein and coronary artery thrombosis: effect of plasminogen activator inhibitor type 1 neutralizing monoclonal antibody. Circulation. 1995;91:1175-1181.
Carmeliet P, Stassen JM, Schoonjans I, Ream B, van den Oord JJ, De Mol M, Mulligan RCM, Collen D. Plasminogen activator inhibitor-1 gene-deficient mice, II: effects on hemostasis, thrombosis and thrombolysis. J Clin Invest. 1993;92:2756-2760.
Orita M, Iwahana H, Kanazawa H, Hayashi K, Sekiya T. Detection of polymorphisms of human DNA by gel electrophoresis as single-strand conformation polymorphisms. Proc Natl Acad Sci U S A. 1989;86:2766-2770.
Condie A, Eeles R, Borresen AL, Coles C, Cooper C, Prosser J. Detection of point mutations in the p53 gene: comparison of single-strand conformation polymorphism, constant denaturant gel electrophoresis, and hydroxylamine and osmium tetroxide techniques. Hum Mutat. 1993;2:58-66.