Apolipoprotein E Polymorphism Is Associated with Both Number of Diseased Vessels and Extent of Coronary Artery Disease in Czech Patients with Cad

Aims. The impact of ApoE polymorphism on angiographic parameters was assessed in patients referred for coronary angiography. Methods. Elective coronary angiography was performed in 671 subjects (525 men, 146 women, mean age 60±10 years) with symptoms of ischemic heart disease. The patients were divided into: no CAD group (smooth coronary vessels, n=83), one-vessel (n=155), two-vessel (n=170) and three-vessel disease (n=196). Patients with stenoses 0-50% were excluded. Within patients with CAD, we evaluated overall extent of CAD measured by the number of stenotic segments according to AHA (1 segment vs. 2–3 vs. ≥4), and the severity of the most serious stenosis (in percent). ApoE genotype was determined using real-time PCR. Results. The frequency of ε2/ε3 genotype (n=56) was lower in the three-vessel disease group compared to one-vessel disease (OR=0.25, P=0.0019), two-vessel disease (OR=0.31, P=0.0114) or no CAD group (OR=0.24, P=0.0057). Frequency of ε2/ε3 decreased with the number of affected segments (1 vs. ≥4: OR=0.35, P=0.0143). The ε3/ε4+ε4/ε4 genotypes (n=123) were more frequent in CAD patients altogether compared with no CAD group (OR=2.30, P=0.019), while no impact of the ε4 allele on angiographic parameters within the CAD patients was detected. In ε2/ε3 carriers with CAD, lower LDL-cholesterol, total cholesterol and lower use of lipid-lowering drugs were observed. Conclusions. The results show predominantly focal form of CAD in patients with ε2/ε3 genotype. Lower LDL-cholesterol and total cholesterol may play the key role, although other contributing factors are discussed.


INTRODUCTION
Coronary artery disease (CAD) is one of the leading causes of both morbidity and mortality in the industrialized world, being itself responsible for 24% of all deaths in the Czech Republic 1 .Numerous environmental and genetic factors have been described, including genes participating in lipid metabolism.
Apolipoprotein E (ApoE) is an important part of several types of lipoprotein particles, including chylomicrons, VLDL, IDL and ApoE-HDL.It is produced by hepatocytes in the liver, by macrophages in peripheral tissues and locally by glial cells in the brain 2 .The protein consists of 299 amino-acids.Its function is essential for lipid metabolism, at both systemic and local level.Among others, ApoE is a ligand of LDL (ApoE/ApoB) receptor (LDL-R) (ref. 3,4), contributing to chylomicron remnant and VLDL clearance 5 , and having numerous functions in vessel wall [reviewed in (ref. 6)].
ApoE is produced by macrophages in diseased arterial wall during its inflammatory reaction, where it helps to carry away cholesterol in locally formed lipoproteins and reduce atherosclerotic plaque 7,8 .This effect can be also mediated by hepatic ApoE, as was proven in mice lacking macrophage-derived ApoE, and is responsible for anti-aterogenic action even without lowering plasma cholesterol levels [9][10][11] .The protein also induce the synthesis of NO in vessel wall 12 and has numerous anti-inflammatory effects 13,14 ApoE is polymorphic, with three isoforms common in general population: E2, E3 and E4.The three isoforms differ in two amino-acids at residues 112 and 158 of the protein.While E4 has arginine at both sites, E3 has cysteine at the first site and arginine at the second one and E2 has cysteine at both sites 15,16 .They are coded by three co-dominant alleles of apolipoprotein E gene (APOE) -ε2, ε3 and ε4, which give rise to 6 possible genotypes: ε2/ ε2, ε2/ε3, ε2/ε4, ε3/ε3, ε3/ε4 and ε4/ε4.

Aim of the study
We assessed the impact of ApoE polymorphism on the severity of CAD in patients referred for coronary angiography.CAD was defined as at least 50% stenosis of any segment of the coronary arteries.One, two and three-vessel disease group were defined by any major segment of LAD, LCx or RCA, affected by ≥50% stenosis.Patients with no narrowing of coronary arteries were used as control group, those with <50% narrowing were excluded, as well as the patients with suspected vasospastic angina pectoris and patients after heart transplantation.
In subsequent analysis of patients with CAD, designed as case-case comparisons, we also focused on the association of the number of stenotic segments with ApoE genotypes, to determine the extent of CAD.The coronary artery tree was divided into segments according to the 16-segment scheme of the American Heart Association.Then, we divided the patients with CAD into 1 stenotic segment group (focal form of CAD), 2 or 3 stenotic segments group (medium form of CAD), and ≥4 stenotic segments group (diffuse form of CAD) with similar numbers of patients.
In patients with CAD, we also investigated an impact of ApoE genotype on the severity of the most serious stenosis in coronary artery tree, expressed in percent of luminal obstruction in two dimensions.
The secondary aim was investigation of the levels of blood lipids and lipid-lowering medication, to determine possible effects of the ApoE isoform on blood lipids.The data on treatment were collected for all subjects, while the lipid spectrum was determined only in 382 (see results).

Coronary artery angiography
Elective coronary angiography was performed in 671 consecutive patients of the 1 st Dep.Internal Med.& Cardioangiology St. Anne's University Hospital in Brno, (525 men and 146 women, age median (lower-upper quartile) 58.5 years (52.9-67.9))with the suspicion for ischemic heart disease (IHD).All subjects were of Caucasian descent.Indications for coronary angiography were IHD-like chest pain and/or results from non-invasive examinations supporting the diagnosis.The participants signed written consent prior to being taken into the study.Coronary angiograms were assessed by four experienced  invasive cardiologists.After the exclusions, we analysed 604 subjects, of whom 521 patients had CAD (155 in onevessel disease group, 170 in two-vessel disease group, 196 in three vessel-disease group).The control group (without any pathology of coronary vessels) was composed of 83 patients (Table 1.).

Laboratory methods
Genomic DNA was purified from peripheral blood leukocytes by the standard method using the phenol-chloroform extraction and the proteinase K digestion of cells.
Two single nucleotide polymorphisms (SNPs) in exon 4 at the positions rs429358 C/T (112 C/R) and rs7412 C/T (158 C/R) were analysed using Real-time polymerase chain reaction (RT-PCR) method monitored by SYBR ® Green (ref. 17).This method was optimized due to the formation of dimers which interfered with the analysis.Each RT-PCR was carried out in a volume 25 μL containing 1x Power SYBR ® Green PCR Master Mix (Applied Biosystems), 0.15 μM of each primer and 50 ng of genomic DNA.The lenght of the product was 173bp.Four allele specific primers were combined in three reaction mixtures (Table 2,3).The amplification conditions were performed according to Calero et al. (ref. 17), a Real Time PCR System (96-well format, ABI Prism 7000 Sequence Detection System, Applied Biosystems) was used.The real-time PCR took six following steps (Table 4).
The genotyping was performed without prior knowledge of the subject´s status.

Statistical analysis
Statistical analysis was performed using STATISTICA software (StatSoft, version 11).Two-tailed Fisher exact test was used for categorical data, Mann-Whitney U-test for continuous data.Bonferroni test was used for multiple testing corrections, when comparing different groups of subjects.One-way ANOVA with Fisher's LSD post-test and t-tests were used for the comparison of blood lipids concentrations (HDL-cholesterol and triglycerides logtransformed to fit the normal distribution).Logistic re- gression was used for odds ratio and confidence intervals estimation.
In the statistical analyses, the ε3/ε4 and ε4/ε4 were included in one group, because low number of ε4/ε4 carriers (n=7) did not allow effective statistical analysis.The ε2/ε3 carriers were, however, evaluated separately, because of large variability of both laboratory values and clinical expression in ε2/ε2 homozygotes 18,19 .
The ε3/ε4 and ε4/ε4 genotypes were significantly more abundant in patients with CAD altogether compared with the no CAD group (OR=2.30,95% CI 1.12-4.74,P=0.019 for ε3/ε4+ε4/ε4).However, no significant difference between the one-vessel, two-vessel and three-vessel disease group was observed.
In a subanalysis of the CAD group according to the number of stenotic segments, we could evaluate whether ApoE polymorphism has also an impact on extent of CAD in the patients with CAD (Table 6, Fig. 2).
No association between the severity of the most serious stenosis and ApoE genotype in patients with CAD was found, although there was an insignificant trend with ε3/ε4+ε4/ε4 genotypes being associated to more severe stenosis (P=0.076).* The frequency of ε2/ε3 genotype in three-vessel disease group was lower compared to one-vessel disease group (OR=0.25,95% CI 0.10-0.61,P=0.0019, P corr =0.011) and also compared to no CAD group (OR=0.24,95% CI 0.09-0.65,P=0.0057, P corr =0.034).Concerning the lipid spectrum, we found that subjects with ε2/ε3 in the study showed lower total cholesterol compared to ε3/ε3 (P=0.023) and ε3/ε4+ε4/ε4 (P=0.020), and lower LDL-cholesterol compared to the same groups (ε2/ε3 vs. ε3/ε3: P=0.011 vs., ε2/ε3 vs. ε3/ ε4+ε4/ε4: P=0.012, respectively) (Table 7).ε2/ε3 patients compared with the rest of our study group had lower total cholesterol (ε2/ε3 vs. other ApoE genotypes: P=0.039 for patients with CAD, P=0.014 for total sample) and LDLcholesterol (ε2/ε3 vs. other ApoE genotypes: P=0.039 for patients with CAD, P=0.006 for total sample) also when only CAD patients were assessed.This effect was apparent despite the fact that ε2/ε3 carriers were less likely to use lipid-lowering drugs, both among CAD patients and the total sample (P=0.013 for patients with CAD, P=0.005 for total sample), while no difference was observed in ε3/ ε4+ε4/ε4 concerning medication.
This study provides evidence, that ApoE ε2/ε3 genotype is associated not only with lower number of coronary arteries affected, but also with overall less extended form of coronary atherosclerosis, measured by the number of segments affected.The ApoE ε3/ε4+ε4/ε4 seems to be associated with CAD onset rather than extent.The lower concentration of LDL-cholesterol and total cholesterol is apparent in ε2/ε3 subjects, although lipid-lowering treatment is less likely to be used in this group.

DISCUSSION
The association of ApoE to CAD has been described in many studies and meta-analyses.Compared to the most abundant ε3/ε3 genotype, whose effects are considered neutral 20 , the ancestral 21,22 ε4 allele has been found as a risk factor of CAD (ref. 20,23,24), with previous studies giving it an important role 25 .Considered as "thrifty allele" 21 , the allele has been associated with elevated total cholesterol and low density lipoprotein cholesterol level 26 , which we failed to confirm in our study.Its risk potential for the development of CAD is commonly accepted, although negative effects on the development of atherosclerosis in heterozygous genotypes appear to be milder in recent large studies and meta-analyses, and possible publication bias is being discussed [27][28][29][30] .
The view of the role of ε2 allele has been evolving across the decades.The homozygosis for ε2 is crucial for the development of familial hyperlipoproteinemia type III (FH III), a disorder of lipid metabolism manifested by premature atherosclerosis 5,18,31,32 , which was historically the first disease linked to the polymorphism of ApoE (ref. 33).However, the disease developes only under certain conditions, as is hyperinsulinemia, excessive caloric intake or male sex, while most of ε2/ε2 homozygotes remain normolipidemic or hypolipidemic 18,34 .Generally, large variability in serum lipids concentration is typical for ε2/ ε2 (ref. 19,27).In our study group, there was only one ε2/ε2 homozygote with completely determined lipid spectrum, a 68 year old patient showing lipid spectrum disorder with elevated cholesterol, LDL-cholesterol and triglycerides.Therefore we cannot confirm or reject the previous results.
In heterozygous genotypes ε2/ε3 and ε2/ε4, the ε2 allele is considered protective against CAD (ref. 19,27,30,35).The most apparent systemic effect is the lowering of LDL-C and total cholesterol in ε2/ε3 (ref. 26,27).Those characteristics, though not in that apparent way, seem to also take place in ε2/ε4 heterozygotes, where the opposite effects of E2 and E4 isoforms are present 27 .
However, few studies have focused on the severity of CAD and length of atherosclerotic lesions 36,37 .The results of those studies are controversial however.Some found no correlation between the severity of CAD and ApoE polymorphism 38,39 , while others established deleterious effect of ApoE4 isoform and protective effect of E2 isoform on the number of affected coronary vessels 36,37 in various populations.The severity of stenoses has been found to follow the same manner in male Chinese 37 .
The overall extent of coronary atherosclerosis has been investigated mainly in autopsy studies that found ε3/ ε4 genotype to be a risk factor for atherosclerosis extent, while no association in autopsy studies has been proven in ε2/ε3 (ref. 40,41).Low overall numbers of ε2/ε3 carriers in those studies are possible explanation for the fact, that no significant changes in a frequency of ε2/ε3 were observed.The in-vivo Brazilian study 42 found no association of ApoE genotype and extent of atheromatosis.Our study, in contrast, suggest an effect of ε2/ε3 genotype on lower extent of coronary artery disease, while it failed to prove any effect of ε4 allele.
The influence of ApoE polymorphism on CAD can be largely attributed to the differential effect on lipid spectrum, as was proven in a large prospective study by Ward et al. 30 , on 22169 subjects.The study showed expected protective effect of ε2 allele and mildly deleterious effect of ε4 allele (which was of borderline significance), while it failed to show any significant effect after adjustment for LDL-C/HDL-C ratio.
Other, lipid-independent, mechanisms also exist.Concerning the systemic inflammatory response, the antiinflammatory effect of ApoE decreases in the sequence E2>E3>E4 [for review, see (ref. 43)].Interestingly, concentration of serum CRP is isoform-specific in the opposite manner (E2>E3>E4) (ref. 28,44), suggesting either some compensatory mechanism or just indirect involvement of CRP concentration in atherosclerosis.
Differential impact on the induction of endothelial nitric oxide synthase (eNOS) has also been described, eNOS being induced by isoforms E3>E2>E4 (ref. 12).This order does not correspond with prevalence of CAD (ε2/ ε3<ε3/ε3<ε3/ε4<ε4/ε4) (ref. 27).The differential induction of eNOS might enhance the development of narrowing due to smooth muscle proliferation in the place of already formed lesions in ε2 and ε4 carriers, while their overall number and extent might be lower in ε2 carriers because of their lower cholesterol and LDL-C.The eventual significance of this differential induction could be potentially demonstrated by combinatory effects of ApoE polymorphism and eNOS polymorphisms with known impact on CAD development, such as 4a/b (ref. 45,46).
The lower use of lipid-lowering treatment among subjects with ε2/ε3 has been already described 47 and is attributable to lower total cholesterol and LDL.The best response to statins and fibrates for lipid spectrum is present in the ε2 carriers (ref. 48,49), but the ApoE genotype does not seem to influence their effect on the progression of CAD (ref. 50,51).Long-term prospective studies are needed for further evaluation.
Frequencies of ApoE genotypes in our group of CAD patients were, in general, consistent with other studies referring to ApoE and CAD performed in Central European populations including Czech 28,39,52 .The strengths and limitations of our control group are discussed further.

Limitations and strengths
The control group was recruited from selected patients referred for coronary angiography.Because of the strict conditions for admission to coronary angiography, the number of control subjects is much lower than patients with CAD.The genotype frequencies do not have to correspond with those present in healthy persons.Among others, possible presence of undiagnosed vasospastic angina (VSA) might have influenced the results.ApoE has been found to influence VSA occurrence in patients with smooth coronary arteries, with ε4 allele being significantly less frequent in subjects with VSA compared to controls 53 .Although subjects with apparent VSA were excluded from our study, no specific test was used for the systematic search of VSA.
On the other hand, the way we selected the control group may also be regarded as strength.It may not be possible to distinguish patients with asymptomatic CAD or insignificant atherosclerosis in general population.A certain degree of atherosclerosis is not rare even in populations aged <20 years and is present in up to 85% in subjects aged ≥50 years 54,55 , which limits the use of asymptomatic randomly selected subjects as controls.Our control group consisted of individuals without any sign of atherosclerosis in the coronary angiograms, which eliminates any such undiagnosed patients or patients with mild forms of CAD.It is of interest, that ApoE ε3/ ε4+ε4/ε4 showed a trend to higher frequency in a group excluded for insignificant atherosclerosis compared to no CAD group (P =0.033, insignificant after Bonferroni correction).The subjects with insignificant atherosclerosis would probably remain inseparable from no CAD group if we used random population sample.An ideal, but unrealistic, way would be to perform coronary angiography in a sample of apparently healthy volunteers.

CONCLUSION
We revealed a significantly lower frequency of ε2/ε3 genotype in patients with three-vessel disease and more diffuse form of coronary artery disease.The subjects with ε2/ε3 differed in their lipid spectrum (with lower total cholesterol and LDL-cholesterol), despite being less likely to use lipid-lowering drugs.Lower extent of CAD could be thus explained by better lipid profile.

Fig. 1 .
Fig. 1.Frequency of genotypes in one-, two-and three-vessel disease and in no CAD subjects.The graph shows significantly lower frequency of ε3/ε3 genotype in three-vessel disease group compared to one-vessel disease and control group.

Fig. 2 .
Fig. 2. Frequency of genotypes and the extent of CAD.The graph shows a decrease of ε3/ε3 genotype frequency together wiht an increase of CAD extent.

Table 1 .
Characteristics of the subjects.
All continuous variables with normal distribution are described as mean±SD, those with non-normal distribution are marked * and the values represent median (lower-upper quartile).

Table 2 .
Primers for real-time PCR.

Table 5 .
Frequency of genotypes in one-, two-and three-vessel disease and in no CAD subjects.

Table 6 .
Frequency of ApoE genotypes and the extent of CAD.

Table 7 .
Blood lipids in the study group according to genotypes.All continuous variables with normal distribution are described as mean±SD, those with non-normal distribution are marked * and the values represent median (lower-upper quartile).The value of total cholesterol marked † was lower in ε2/ε3 compared to ε3/ε3 (P=0.023) and to ε3/ε4+ε4/ε4 (P=0.020).The value of LDL-cholesterol marked † † was lower in ε2/ε3 compared to ε3/ε3 (P=0.011) and to ε3/ε4+ε4/ε4 (P=0.012).Total cholesterol in ε2/ε3 vs. all other ApoE genotypes: P=0.039 for patients with CAD, P=0.014 for total sample; LDL-cholesterol in ε2/ε3 vs. all other ApoE genotypes: P=0.039 for patients with CAD, P=0.006 for total sample.