Introduction

Endothelial nitric oxide synthase (eNOS, encoded by the NOS3 gene) synthesizes NO from l-arginine and molecular oxygen in vascular endothelial cells (Kincl et al. 2009). Angiotensin-converting enzyme (ACE), present on the surface of vascular endothelial cells, generates the potent vasoconstrictor angiotensin II from angiotensin I and inactivates the vasodilator bradykinin (Erdös 1990). Angiotensin II modulates NO synthesis in cardiovascular tissue, and NO modulates the action of angiotensin II (Dubey et al. 1995; Nakagami et al. 1999). Many polymorphisms located in the ACE and NOS3 genes have been reported to play a major role in the pathogenesis of coronary artery disease (CAD) and related outcomes (Cambien et al. 1992; Yoshimura et al. 2000; Bor-Kucukatay et al. 2010; Hamelin et al. 2011). The −T786C polymorphism in the NOS3 gene causes a reduction of promoter activity and has been reported as a risk factor for coronary spasm in a Japanese population (Nakayama et al. 1999). The Glu298Asp polymorphism has been associated with reduced basal NO production (Veldman et al. 2002) and has been linked to the risk for CAD (Hingorani et al. 1999).

A polymorphic variant in intron 16 (ACE-I/D) of the ACE gene, characterized by an insertion (I) or deletion (D) of a repeat sequence of 287 noncoding base pairs (Zintzaras et al. 2008), has been shown to be associated with hypertension and other cardiovascular risk factors (Rigat et al. 1990). The ACE4 polymorphism in the 5′ UTR has been considered as a tag SNP in African population (Keavney et al. 1998; Zhu et al. 2001), a result that has also been established by a study of a Tunisian population (Rebai et al. 2006). Few studies have addressed the relationship between NOS3 and ACE gene variants and the pathogenesis of CAD (Nakagami et al. 1999; Alvarez et al. 2001). In this work, we report a case–control study in southern Tunisia, investigating the association between CAD and four SNPs: T-786C (rs11771443) and Glu298Asp (rs1799983) polymorphisms in the NOS3 gene, and ACE-I/D (rs4340) and ACE4 (rs4291) polymorphisms in the ACE gene.

Materials and Methods

Study Subjects

We enrolled 249 unrelated patients (185 men and 64 women, 38–81 years old, mean age 57.07 ± 10.6 years) who were diagnosed with CAD by the Cardiology Service of the Hedi Chaker University Hospital of Sfax, Tunisia, from May 2007 to December 2009. Angiography confirmed that each patient had stenosis >50 % of at least one major coronary artery. As a routine procedure, an informed written consent was obtained from all patients.

The control group consisted of 295 healthy unrelated volunteers (202 men and 93 women, 36–69 years old, mean age 55.2 ± 11.7 years) from the staff of the Sfax Center of Biotechnology and blood donors of the Sfax Center of Transfusion. They had no history of vascular disease. Cases and controls were matched by gender, age, and ethnicity.

Biochemical Analysis

The serum concentrations of glucose, triglycerides, total cholesterol, creatinine, urea, uric acid, and blood major ionic constituents (K+, Cl−, and Na+) were measured by the standard methods used in the clinical laboratory of the hospital.

DNA Analysis

Genomic DNA was extracted using the standard phenol/chloroform method from EDTA anticoagulated peripheral blood samples (Marcadet et al. 1987).

The ACE4, −T786C, and Glu298Asp polymorphisms were typed by PCR amplification, followed by restriction enzyme digestion (PCR–RFLP), with the classical PCR for −T786C and Glu298Asp (using the conditions described by Fatini et al. 2004) and nested PCR for ACE4 (conditions of Zhu et al. 2001). PCR products were digested with the appropriate enzymes and run on agarose gels (4 %). For the ACE-I/D polymorphism, no digestion was needed. This analysis included an additional PCR to confirm the absence of an I allele in DD individuals. The ACE alleles were visualized as fragments of 490 bp (I) and 190 bp (D) (Shanmugan et al. 1993).

Statistical Analysis

All statistical analyses were performed using SPSS version 13.0 (Chicago, USA). Haplotypes were inferred using Phase version 2.0.2 (Stephens et al. 2001) to estimate the haplotype frequencies in both groups (patients and controls). The differences in allele, genotype, and haplotype frequencies between the groups were tested by chi-square or Student’s tests wherever appropriate. Hardy–Weinberg equilibrium was tested using the Genetic Data Analyses program, version 1.1 (Weir 1996). A logistic regression analysis was performed to determine independent predictors for CAD. One-way ANOVA was used to analyze the relationship between genotypes and the general characteristics and severity of CAD. A value of p < 0.05 was considered the cutoff for significance.

Results

Genotype, Allele, and Haplotype Frequencies

All genotype distributions of the polymorphisms studied were in agreement with Hardy–Weinberg expectation (p > 0.05). Genotype, allele, and haplotype frequencies were similar between patients and controls (Table 1).

Table 1 Frequency of ACE and NOS3 gene SNPs in a Tunisian population
Full size table

Risk Factor for CAD

Binary logistic regression was used to test the association of the diseases (dyslipidemia or CAD) with SNPs after adjusting for confounding factors (age, sex, smoking, hypertension, and body mass index). No significant association was found for the four polymorphisms (p > 0.05), but hypertension was identified as an acquired risk factor of CAD (p = 0.034) and also of dyslipidemia (p < 0.001). In addition, smoking was identified as a risk factor only for CAD (p < 0.001). Furthermore, a multivariate analysis using one-way ANOVA between patient groups showed no significant association between the four polymorphisms and any of the clinical and biological parameters (p > 0.05; data not shown).

Association Between SNPs, Clinical Characteristics, and CAD Severity

Patients with CAD were classified into three subgroups (G1, G2, and G3) according to the number of affected coronary arteries. The calculation of allele, genotype, and haplotype frequencies revealed no association with CAD severity (data not shown); however, smoking habit (p = 0.008), dyslipidemia (p = 0.028), type 2 diabetes (p = 0.006), and increased rate of glucose (p = 0.033) were found to differ among the three groups of patients and to correlate with increased risk of CAD severity (Table 2).

Table 2 Characteristics of patients and their association with CAD severity
Full size table

Discussion

In the present study, we examined the possibility of association between CAD and four polymorphisms in a southern Tunisian sample: ACE4 and ACE-I/D polymorphism in the ACE gene, and −T786C and Glu298Asp in the NOS3 gene with CAD. We found no association, and this finding persisted after adjusting for several potential confounding factors.

Regarding the Glu298Asp polymorphism in the NOS3 gene, no association has been reported with CAD in Asian and European populations (Aras et al. 2002; Fatini et al. 2004; Kim et al. 2007; Guldiken et al. 2009), which agrees with our finding. This polymorphism was not associated with hypertension in the Tunisian population (Sediri et al. 2010). In contrast, previous studies showed that this polymorphism seemed to be significantly and independently associated with the occurrence and severity of CAD in Italian and Japanese populations (Yoshimura et al. 1998; Ghilardi et al. 2002; Colombo et al. 2003). These contradictory results might lie in the ethnic origins of the populations studied or in differences in selection criteria and sample sizes.

Next, we found no association between the polymorphism −T786C in the NOS3 gene and CAD. The results were the same in studies involving European and Australian populations, but the relation between this SNP and CAD remained controversial (Marroni et al. 2005). The meta-analysis reported by Casas et al. (2004) for the Glu298Asp and −T786C polymorphisms showed a difference in allelic frequencies for Asians versus non-Asians for both polymorphisms. These interethnic differences might in part explain the ethnic disparities in NO bioavailability, cardiovascular risk, and response to drugs (Marroni et al. 2005).

No association with CAD was found for haplotypes of the NOS3 or the ACE gene in our population. In contrast, Sandrim et al. (2007) reported that the C-Glu haplotype decreased the risk of developing hypertension and was associated with higher nitrite/nitrate (NO x ) levels in hypertensive patients, whether combined or not with type 2 diabetes mellitus, although individual NOS3 polymorphisms did not have significant effects. Moreover, this same specific haplotype was involved in the modulation of NO formation (Metzger et al. 2005, 2007, 2011) in healthy Caucasian subjects. Our study included more enrolled subjects, which might have increased the power of our study.

The patients who had dyslipidemia were taking statins, which increased NOS3 expression and up-regulated NO formation (Lacchini et al. 2010). No significant association was found between dyslipidemia and the SNPs using binary logistic regression. On the other hand, several studies carried out on Caucasians have shown that only healthy subjects with the CC genotype for T-786C receiving treatment with a low dose of atorvastatin for 2 weeks had augmented NO availability, produced antioxidant (Nagassaki et al. 2006) and anti-inflammatory (Souza-Costa et al. 2007) effects, and reduced membrane fluidity of erythrocytes (Nagassaki et al. 2009). These studies included only a small number (30) of healthy male subjects, which might have limited power to detect the difference between groups and may restrict the conclusions to this specific population.

Regarding the ACE gene polymorphism, we found no association for ACE4 with CAD. Zhu et al. (2001), however, found a positive correlation between blood pressure and plasma concentration in an African population. This result was expected because allele frequencies in our population are different from those of Africans and Europeans (El Moncer et al. 2010). For the ACE-I/D polymorphism, a recent meta-analysis conducted by Zintzaras et al. (2008), including 118 studies (43,733 cases with CAD and 82,606 controls), reported a significant association for European populations (odds ratio 1.25, DD vs. II). Furthermore, Dzimiri et al. (2000) stated that no association has been reported to date in Arab or North African populations. We can explain this finding by the heterogeneous genetic profile of the Tunisian population, which is characterized by important emetic exchanges throughout history and frequent migration around the Mediterranean Sea (Maalej et al. 2004). The authors ranked the Tunisian population between the Sub-Saharan African and the Caucasian populations. Another explanation of our result may be the presence of linkage disequilibrium between the ACE-I/D polymorphism and other polymorphisms in this region. Indeed, Keavney et al. (1998) suggested the presence of functional polymorphisms located between intron 18 and the 3′ UTR and excluded the ACE-I/D marker within intron 16 in this region.

Furthermore, conventional risk factors were highly prevalent in patients, as expected (Vogel and Motulsky 1997; Wilson and Culletton 1998). These results were further supported by regression analysis, which demonstrated, after correction, that only smoking could be identified as a dependent acquired risk factor for CAD. Hypertension was identified as a risk factor for CAD and dyslipidemia. Our study lacks the assessment of NO formation in CAD patients and controls, which could improve our finding.