|Home | About | Journals | Submit | Contact Us | Français|
The majority of acute coronary syndrome (ACS) cases cannot be explained by the analysis of commonly recognized risk factors; thus, the analysis of possible genetic predispositions is of interest. The genes for connexin-37, stromelysin-1, plasminogen activator-inhibitor type 1 (PAI-1) and lymphotoxin-alpha are among many presently known candidate genes that are associated with risk factors for ACS.
To identify the potential impact of the functional variants of connexin-37, stromelysin-1, PAI-1 and lymphotoxin-alpha on ACS in a Caucasian Czech population.
A total of 1399 consecutive patients (1016 men and 383 women) with ACS from five coronary care units located in Prague (Czech Republic) were analyzed; a representative sample of 2559 healthy individuals (1191 men and 1368 women) were also genotyped and served as controls.
The gene variants analyzed were not significantly associated with the prevalence of ACS or the classical risk factors of ACS development such as high plasma lipid levels, hypertension, diabetes, high body mass index or smoking.
In a Caucasian Czech population sample, genetic variants of connexin-37, stromelysin-1, PAI-1 and lymphotoxin-alpha were not significantly associated with a predisposition toward ACS.
A substantial portion of coronary artery disease (CAD) events are explained by the commonly accepted ‘classical’ risk factors of cardiovascular disease such as hypertension, dys-lipidemia, obesity, smoking and diabetes. However, one-third of cardiovascular disease cases cannot be characterized by these generally accepted risk factor criteria and, despite intensive research over the past decades, no major advances in the detection of additional risk factors have been made. Furthermore, the current worldwide prevalence of CAD is increasing and is expected to escalate even further in the future (1).
Molecular genetic methods offer an alternative to using the classical risk factors for assessing CAD risk. Because of the high heritability of CAD, it may be possible for this predisposition to be detected by the identification of risk variants in the candidate genes believed to be associated with CAD. Further advances in the molecular genetic approach to CAD will contribute to better understanding of the differences between individuals, and may offer more effective treatment for those detected to be at risk.
The complexity of CAD, however, makes the application of molecular methods somewhat more difficult, but two extreme models of CAD have emerged (2). The ‘polygenic’ model proposes that a large number of different alleles with small effects at different genes interact with one another and with the environment to cause cardiovascular disease. The ‘monogenic’ model proposes that rare alleles with substantial effects at a large number of genes cause cardiovascular disease. In atherosclerosis, a combination of these two models with strong additional environmental influence (eg, smoking, high fat intake, lack of physical activity and others) is believed to be involved. Cardiovascular disease was calculated to have a significant heritability component of approximately 60% in men and approximately 40% in women (3), and that the risk of premature cardiac death was eightfold higher in men and 15-fold higher in women whose siblings died as a result of heart attack before 55 or 65 years of age, respectively (4).
Despite the known rare mutations, the majority of myocardial infarction (MI) risk is under polygenic control, with many variants in a few genes contributing to the total effect. The estimated effect of each individual single nucleotide polymorphism (SNP) is estimated to be relatively low, yet significant. Because atherosclerosis and ischemic heart disease are very complex and nonhomogeneous diseases, recent estimates suggest that as many as 500 to 800 genes influence the patho-physiological processes from the onset of atherosclerosis to its clinical manifestation. Consequently, thousands of articles in the field of genetic predisposition to cardiovascular disease have been published – the results of which are difficult to replicate due to low statistical power, usually owing to a small number of participants and ethnic differences.
Two interesting genetic studies (12,13) from Japan identified four genes of considerable interest that had a highly significant effect on CAD development. These are the genes for connexin-37, a gap junction protein involved in cell-to-cell communication; stromelysin-1, a matrix metalloproteinase-3 involved in the destabilization of atherosclerotic plaques; plasminogen activator-inhibitor type 1 (PAI-1), a protein responsible for prothrombogenic status of plasma; and the gene for lymphotoxin-alpha – also known as tumour necrosis factor-beta – a toxin produced by lymphocytes associated with inflammatory processes in artery walls leading to the initiation, progression and destabilization of atherosclerotic plaques.
The significant differences in genetic predisposition toward CAD, mainly among different ethnic groups, prompted us to analyze the putative association between SNPs within these four genes and ACS in a large group of Caucasian Czech patients and healthy controls.
A total of 1686 patients younger than 65 years (men) or 75 years (women) of age hospitalized in five participating coronary units (Clinic of Cardiology, Institute for Clinical and Experimental Medicine; 2nd Department of Internal Medicine, General Teaching Hospital; Department of Cardiology, Teaching Hospital Motol; Department of Cardiology, Homolka Hospital; and the Cardiocenter, Department of Cardiology, University Hospital Královské Vinohrady) were included in the present study (14). These facilities cover almost all areas of health care in Prague (Czech Republic) for ACS (eg, acute MI, minimal myocardial lesion and patients with subacute MI). The only exclusion criteria were age of 65 years and older for men, 75 years and older for women, and refusal to participate in the study. Of the 1686 patients, detailed genetic association analyses were performed for 1399 individuals (1016 men and 383 women) for which complete biochemical and anthropometrical data were available.
The control groups (1191 men and 1368 women [response rate 84%]) were selected from 1% of a representative three-year cohort sample of a Caucasian Czech population (from the districts of Benesov, Prague-east, Cheb, Chrudim, Jindrichuv Hradec, Pardubice, Kromeriz, Litomerice and Pilsen City). The individuals were recruited in nine districts from 1997 to 1998 and re-invited in 2000/2001 according to the WHO protocol (“MONICA Project”. Manual WHO/MNC 82.2, November 1983). The local ethics committee approved the study design and written, informed consent was obtained from the participants. Individuals in the patient and control groups completed a standard questionnaire regarding the presence of traditional cardiovascular risk factors including family and personal history. The characteristics of the study participants are summarized in Table 1.
Patient and control DNA was isolated as previously described (15). Individual SNPs (4G/5G polymorphism at −675 of the PAI-1 promoter; A252→G of the lymphotoxin-alpha gene, C1019→T in the connexin-37 gene; and the 5A/6A variant at −1171 bp of the stromelysin-1 gene) were genotyped using a high-throughput microplate array diagonal gel electrophoresis methodology (16) (Table 2). The lipoprotein parameters were measured at the WHO Regional Lipid Reference Centre, IKEM (Prague) on a COBAS MIRA auto-analyzer (Roche, USA).
The Hardy-Weinberg test was used to confirm the independent segregation of the alleles of individual genotypes (www.tufts.edu/~mcourt01/Documents/Court%20lab%20-%20HW%20calculator.xls). The χ2 test was used to test for differences between the genotype frequencies of the participating groups (www.physics.csbsju.edu/cgi-bin/stats/contingency_form.sh?nrow=2&ncolumn=3). ANOVA was used for additional statistical analysis.
For genotype/allelic differences between the groups, P<0.05 was considered to be statistically significant. Because of the high number of analyses performed, P<0.01 was considered to be significant in the analysis of the putative association between individual SNPs and the classical risk factors.
The basic characteristics of the study groups are summarized in Table 1. As expected, patients (both men and women) were older and had a higher prevalence of smoking, diabetes and hypertension than the control groups. However, the mean body mass index of the male patients was similar to the mean body mass index of the control group, and plasma cholesterol levels were lower in ACS patients. This was due to the fact that approximately 20% of the patients (in contrast to approximately 8% of the controls) were on lipid-lowering drugs (almost all on statins) at the time of ACS.
The distribution of all analyzed SNPs were similar to the frequencies described in other Caucasian populations (Table 3). Overall, the call rates for individual genotypes varied between 93.7% (connexin-37 in male control groups) and 100% (PAI-1 variant in female control groups). With exception of the PAI-1 variant (borderline, P=0.03 and just for female controls), the distributions of individual genotypes were within the Hardy-Weinberg equilibrium. The frequency of the genotypes and alleles containing the analyzed SNPs were not significantly different between MI patients and healthy controls (all P≥0.2) (Table 3).
The present study failed to identify associations between classical MI/CAD risk factors (total cholesterol, triglycerides, diabetes prevalence, blood pressure and glucose) and the four analyzed gene variants, either in patients or controls, or in men and women.
The present large study (investigating approximately 1500 patients and 2500 controls) failed to replicate previous findings of an association between four functional SNPs within the genes for connexin-37, PAI-1, stromelysin-1 and lymphotoxin-alpha with ACS. Furthermore, these variants were not significantly associated with the ‘classical’ risk factors for ACS/CAD development.
Our results contrast with the original study (12) that found the four genes to be important determinants of MI development in a Japanese population; however, later studies (13) on these variants were less clear.
The most controversial results have been published in association with the C1019→T polymorphism in the connexin-37 gene – a variant that has been associated with cardiovascular risk; however, it remains unclear as to which allele carries this risk. Studies conducted in Japan (12,17) showed that the T allele contributed to MI development, especially in high-risk individuals; in contrast, however, studies on Caucasians (18,19) with the same variant showed it to be protective. The present study – and others (20) – did not detect an association between the C1019→T variant and CAD/ACS risk.
The finding of an association between the lymphotoxin-alpha variant and ACS suggests a strong ethnic-specific effect. This variant was originally associated with CAD and later confirmed in Japanese populations (12,13), but not in German (21) or Brazilian (22) populations.
The 4G/5G variant of the PAI-1 gene is among the most analyzed hemostatic genetic polymorphisms in CAD. A large meta-analysis (23) determined the per allele RR for CAD to be significant, but very low (OR 1.06). Despite the large sample size, the present study was inadequately powered to detect such a small effect.
Finally, the available data regarding the stromelysin variant are controversial. The largest study (24) performed to date did not find an association between stromelysin and MI; however, the same variant was associated with CAD.
There are many reasons that could explain the differences in the results of our study with previous findings. First, it is known that the prevalence of MI is much higher in European than in the Japanese populations, possibly due to a difference in the gene effects in the two groups. This is supported by recent findings analyzing thousands of SNPs. Briefly, if a large number of SNPs are analyzed, different SNP patterns emerge, resulting in many distinct ‘genetic fingerprints’ that may have different effects in different populations (25). This may also apply to some and, perhaps most, key individual SNPs. In fact, not all of the thousands of SNPs analyzed are of functional relevance.
Second, gene-environment interactions definitely play an important role. The negative gene effect seen in the Caucasian Czech population may be partially masked by an unfavourable environment. Generally speaking, a large proportion of the Japanese diet consists of fish, which is highly regarded to be a low-risk diet with respect to CAD. In contrast, a large proportion of the typical Czech diet consists of red meat. Indeed, there are published studies underlining the importance of individual SNPs in population analyses that are placed in such a context. For example, the effect of a common variant within the promotor region of the hepatic lipase gene on plasma lipoprotein levels was shown to be dependent on dietary fat intake (26).
Our negative results underline the importance of confirmatory analyses performed on a large number of appropriately selected and characterized groups of patients and controls, especially when considering populations with different ethnicities and lifestyles.
Finally, it must be emphasized that the selection of the genes analyzed in the present study was based on a candidate gene approach, which is supported by functional data and published results. Recently, genome-wide association studies investigating many thousands of SNPs (whose selection was often based merely on the criterion ‘to equally cover the genome’), mostly without known functionality, were analyzed. Interestingly, of the significant SNPs detected and replicated through genome-wide association studies, a substantial portion were localized in new genes, mostly with unknown functions and vice versa, and many candidate genes were not confirmed (but definitely not excluded, and many other SNPs, especially with lower frequency, were not covered by the genotyping chips used).
Both genetic and environmental factors contribute to ACS risk in approximately equal proportions. However, with intensive prescription of some medications (eg, lipid- and blood pressure-lowering drugs), it is likely that future research will reveal an even greater role of genetic predisposition.
We suspect that in the Czech-Slavonic population, genes other than the connexin-37, PAI-1, stromelysin-1 and lymphotoxin-alpha genes (eg, genes for apolipoproteins A5 and E) (27–29) are likely to play a more important role in the detection and development of ACS.
In the future, the detection of CAD risk alleles may lead, in time, to the identification of individuals at risk. When classical risk factors for CAD/ACS (anthropometrical and/or biochemical) are already present, it is clearly too late for primary prevention because the disease is, more or less, already in progress.
This study was supported by research grant No. NR-9093-4/2006 (Internal Grant Agency of the Ministry of Health, Czech Republic).