Plasminogen Activator Inhibitor-1 Promoter 4G/5G Genotype and Plasma Levels in Relation to a History of Myocardial Infarction in Patients Characterized by Coronary Angiography

Subjects

Four hundred fifty-three consecutive Caucasian European patients living in Yorkshire (320 men and 133 women) who were admitted for routine angiography for investigation of chest pain or suspected CAD were recruited from two centers, located in Leeds and Wakefield. Each subject gave informed consent, and the study was approved by the United Leeds Teaching Hospitals (NHS) Trust and Pinderfields Health Trust Research Ethics Committee.

Analytical Methods

Each patient was studied between 7:00 and 10:30 AM after an overnight fast of at least 8 hours. Free-flowing blood samples were taken from an antecubital vein using a 19-gauge butterfly needle. Blood was taken into 0.9% citrate on ice for PAI-1 assay and centrifuged at 2560g and 4°C for 30 minutes before storage at −40°C. Blood was also collected into EDTA for DNA extraction by a salt detergent method as described previously.⁹ Samples were collected for measurement of total cholesterol and TG levels.

Smoking history was determined, and patients were classified into two groups: current smokers, which included those who had ceased smoking in the last 10 years, and nonsmokers, which included those who had stopped smoking more than 10 years previously. BP was measured with subjects lying down and to the nearest 2 mm Hg using the Dynamap automated sphygmomanometer (model 1846 SX/P, version 086, Critikon). The diagnosis of MI was ascertained from patients' hospital records using the WHO criteria of at least 2 of 3 from the electrocardiographic finding of ST-segment elevation of 1 mm in two or more successive leads, typical chest pain longer than 20 minutes' duration, and creatinine kinase increase of more than twice the baseline value. Patients who were reported to have had a history of MI but who did not meet the WHO criteria (n=8) were excluded from the study. In a majority of cases the diagnosis was confirmed by the presence of hypokinetic and akinetic segments at angiography. Patients were recruited into the study irrespective of the time of the MI. BMI was calculated as weight in kilograms divided by the height in meters squared.

Plasma PAI-1 antigen was measured by ELISA (Imulyse, Biopool). Total cholesterol and total TG levels were measured using the Hitachi 747 automatic analyzer (Boehringer Mannheim).

PAI-1 4G/5G promoter genotype was established for each subject by allele-specific polymerase chain reaction amplification of genomic DNA as described previously.¹⁰ Each patient was classified into one of three possible genotypes: 4G/5G, 4G/4G, or 5G/5G. To validate this method of genotyping, control subjects of each genotype as shown by direct sequencing were run with each batch of samples. It was not possible to genotype 19 of the patients, and they were excluded from further analysis.

Coronary Angiography

Angiography was carried out and reported by cardiologists who had no knowledge of patients' genotype or other data. Patients were classified as having normal vessels or single-vessel or multivessel disease on the basis of a ≥50% stenosis in one or more of the three major coronary arteries or their branches.

Statistical Methods

Values for PAI-1 antigen and TG were log-transformed to permit the use of parametric tests. Differences in frequencies of categorical variables (eg, genotype, extent of coronary disease) between groups were assessed by χ² test. Differences in means of continuous variables between male and female subjects were compared by Student's t test. Differences in means of continuous variables between more than two groups were assessed by one-way ANOVA. Differences in continuous variables between groups with adjustment for other factors were assessed by factorial ANOVA. Pearson correlation coefficients were calculated to assess the relationships between the various risk factors for MI and PAI-1 levels. To investigate possible interactions between PAI-1 genotype and other factors, we compared regression slopes between genotype groups, and factorial ANOVA was used to assess the significance of any such interaction as detailed in “Results.” Statistical tests were performed using the SPSS Package for Windows, version 6.1, SPSS Inc.

PAI-1 Genotype and PAI-1 Levels

In keeping with previous studies we found an association between PAI-1 promoter (4G/5G) genotype and PAI-1 levels after adjusting for BMI and TG level, with the highest circulating levels in 4G/4G subjects, intermediate levels in 4G/5G heterozygotes, and lowest levels in 5G/5G subjects, suggesting a codominant gene effect.^{5 6 7 8} We have also repeated the finding of others that the relationship between fasting PAI-1 and TG levels is influenced by PAI-1 promoter genotype, with a steeper regression slope in the 4G/4G group.^{7 23} We also found a similar influence of genotype on the regression slope of PAI-1 on BMI, although in this case the slope was steeper in the 5G/5G group. The interactions of both TG level and BMI with genotype remained both independent and significant in a regression model. Although the interaction of TG level with genotype fits with the finding of higher levels of PAI-1 in the 4G/4G group, it is unclear exactly what the statistical interaction between BMI and the 5G/5G genotype tells us about the regulation of PAI-1. In a similar study in Pima Indians, an interaction between PAI-1 genotype and BMI was observed but with a steeper regression slope in the 4G/4G group.²⁴ Clearly, ethnic differences or chance may account for these findings, and further studies will be required to confirm significant interaction between BMI and genotype in relationship to PAI-1 levels.

Associations With Coronary Atheroma

The proportions in the various angiographic groups were similar to that reported in the ECAT study² and identical to data from the CASS registry.²⁴

This study found the expected relationship between the presence of coronary stenosis with age, BMI, presence of hypertension, cigarette smoking history, and cholesterol levels. However, the relationship between PAI-1 levels and the presence of stenosis in one or more coronary arteries was weak, falling just above standard levels of significance, and there was no trend in PAI-1 levels in relation to the extent of coronary disease. In this latter respect our results are in agreement with those of the larger ECAT study.² To the best of our knowledge this is the first study to investigate the relationship between the extent of coronary atheroma and PAI-1 promoter genotype, and we found no evidence of an association with the frequency of the potentially deleterious 4G/4G genotype, being similar in each angiographic class. This suggests that genotype and thus pre-morbid PAI-1 levels do not influence the generation of coronary atheroma. If this is the case, then the relationship between PAI-1 levels and the presence of coronary atheroma that others have shown may have resulted from the production of PAI-1 in response to or triggered by the presence of atheroma. An alternative explanation for their results is that within the patient group without significant coronary stenosis a number of patients may have had microvascular coronary disease²⁵ and that PAI-1 genotype may influence the development of microvascular disease in the same way as macrovascular disease. However, there are no previous data to support or refute this and the majority of the patients with “clean” coronary arteries are likely to have suffered from either coronary vasospasm or noncardiac chest pain.

Associations With Coronary Thrombosis

As would be expected, a history of MI was more frequent with male sex, a history of cigarette smoking, and an increased number of stenosed coronary arteries, as well as being related to cholesterol levels, BMI, and age. Although there was only a trend to higher levels of PAI-1 in the MI group, we found a significant association between PAI-1 genotype and history of MI that remained after adjusting for the other risk factors. A single measurement of PAI-1 levels is unlikely to reflect lifelong levels because it is an acute phase reactant. Genotype, which does not change, may act as an effective marker for lifelong exposure to differing levels. This could explain why we have been able to demonstrate a relationship between genotype and a history of MI, whereas we were able to show only a trend between PAI-1 levels and a history of MI.

Our findings concur with the results of two small and restricted studies that found the 4G/4G genotype to be more frequent in young male survivors of MI⁶ and in non–insulin-dependent diabetes mellitus patients with a history of ischemic heart disease.¹⁰ On the other hand, our results differ from those of the large four-center ECTIM study in which no association between PAI-1 genotype and MI was found.⁸ The differences between our study and ECTIM are important because although they may reflect the play of chance they may also suggest the way in which PAI-1 genotype affects the processes causing coronary disease.

PAI-1 Genotype, Coronary Atheroma, and Coronary Thrombosis

In the ECTIM study, subjects who had survived MI were compared with control subjects recruited from the general population in the three French centers and Belfast.⁸ Although these subjects were apparently healthy, the frequency of significant, although asymptomatic, coronary disease in this control group is not known. The ECTIM study was not designed to be able to distinguish the influence of genotype on coronary thrombosis from that on atherogenesis. In contrast to this, our study examined both coronary atheroma and thrombosis, finding no relationship between PAI-1 genotype and extent of atheroma but finding an association with previous coronary thrombosis. A hypothesis whereby preexisting coronary atheroma may be necessary for the effect of PAI-1 genotype on coronary thrombosis to become evident, presumably through altered PAI-1 levels, is supported by the finding that the relationship between genotype and MI strengthened when only those patients with significant coronary artery stenosis were examined. Furthermore, this influence of PAI-1 genotype on PAI-1 levels leading to a prothrombotic state could take place at the local level. PAI-1 production is increased at and around atherosclerotic plaques,³ so the rate of both basal and stimulated PAI-1 gene transcriptions may be influenced by PAI-1 promoter genotype through molecular mechanisms being elucidated. This would account for the weak association that we found between history of MI and levels of PAI-1 in the systemic circulation.

In conclusion, our findings suggest that genotype at the PAI-1 promoter polymorphism is an independent risk factor for MI in patients being investigated for chest pain or suspected coronary disease and particularly in subjects known to have significant coronary artery stenosis. Our results indicate that this effect is mediated by involvement of the PAI-1 gene in the pathogenesis of coronary thrombosis, with the greatest risk found in those subjects possessing the 4G/4G genotype. Our findings also support the hypothesis that elevated levels of PAI-1 may predate the development of disease and indicate that PAI-1 genotype may serve as a useful marker of future risk. Prospective longitudinal studies are warranted to confirm these findings.