Clinica Chimica Acta 436 (2014) 11–17

Contents lists available at ScienceDirect

Clinica Chimica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /c l inch im

The influence of apolipoprotein A5 T-1131C and apolipoprotein Ecommon genetic variants on the levels of hemostatic markers indyslipidemic patients

Dalibor Novotny a,⁎, David Karasek b, Helena Vaverkova b, Ludek Slavik c

a Department of Clinical Biochemistry, University Hospital Olomouc, I.P. Pavlova 6, 775 20 Olomouc, Czech Republicb 3rd Department of Internal Medicine, Faculty of Medicine and Dentistry, Palacky University Olomouc and University Hospital Olomouc, I.P. Pavlova 6, 775 20 Olomouc, Czech Republicc Department of Haemato-Oncology, Faculty of Medicine and Dentistry, Palacky University Olomouc and University Hospital Olomouc, I.P. Pavlova 6, 775 20 Olomouc, Czech Republic

Abbreviations: MetS, metabolic syndrome; NCEP, NProgram; ATP, adult treatment panel; SBP, systolic blood presure; BMI, body mass index; CRP, C-reactive protein; TC, toHDLc, high density lipoprotein cholesterol; LDLc, low dnonHDL, TC− HDLc; AI, atherogenic index of plasma (logLp(a), lipoprotein (a); vWF, von Willebrand factor; tPA, tis1, plasminogen activator inhibitor-1; T, C, alleles of Apo AE23, E24, E33, E34 and E44, Apo E genotypes.⁎ Corresponding author. Tel.: +420 585854230.

E-mail address: dalibor.novotny@fnol.cz (D. Novotny)

http://dx.doi.org/10.1016/j.cca.2014.04.0290009-8981/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 16 March 2014Received in revised form 19 April 2014Accepted 21 April 2014Available online 9 May 2014

Keywords:Metabolic syndromeDyslipidemiaApolipoprotein A5Apolipoprotein EGenetic polymorphismTissue plasminogen activator

Objectives: The aim of this study was to evaluate the relationships of the T-1131C (rs662799) polymorphismvariants of apolipoprotein A5 (Apo A5) gene and variants of apolipoprotein E (Apo E) gene common polymorphism(rs429358, rs7412) to selected hemostatic markers.Study design and methods: We examined 590 asymptomatic dyslipidemic patients, subsequently divided intoMetS+ (n = 146) and MetS− (n = 444) groups according to the criteria for identification of the metabolicsyndrome (MetS). We compared variant frequencies and differences in levels of hemostatic markers according toApo A5, Apo E and Apo A5/Apo E common variants.Results: The−1131CApoA5minor variantwas associatedwith elevated tissue plasminogen activator (tPA) in com-parison to TT genotype (p b 0.001), but not in theMetS+ group. The analysis of Apo A5/Apo E common variants inall subjects revealed that the presence of−1131Cminor allele has always been associatedwith higher levels of tPAin comparison with T allele, regardless of Apo E genotype. Also the presence of minor Apo E2 allele led to elevatedtPA concentrations in both T and C carriers. In addition, common −1131C/E2 variant was associated with the

highest tPA levels.Conclusion: We demonstrated a remarkable association especially between the −1131C Apo A5 variant andincreased tPA levels in asymptomatic dyslipidemic patients.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Recent epidemiological and clinical studies support a connectionbetween obesity and thrombosis represented by an increased expressionof the prothrombotic markers and platelet activation. In obese patients,clinical markers of a prothrombotic state might indicate a risk for thedevelopment of complication of metabolic syndrome [1].

Themetabolic syndrome is a commonmetabolic disorder associatedwith an increased risk of type 2 diabetes mellitus and cardiovasculardiseases [2]. The association with elevated risk of atherothrombotic

ational Cholesterol Educationssure; DBP, diastolic blood pres-tal cholesterol; TG, triglycerides;ensity lipoprotein cholesterol;TG/HDLc); Apo, apolipoprotein;sue plasminogen activator; PAI-5 gene at position −1131; E22,

.

cardiovascular events and venous thromboembolism is described aswell [3]. In MetS, enhanced coagulation as well as impaired fibrino-lysis was observed. High levels of fibrinogen and plasminogen activator-inhibitor 1 (PAI-1), togetherwith increased vonWillebrand factor (vWF)and tissue plasminogen activator (tPA) concentrations could reflectendothelial dysfunction [4]. This process is regarded as an early step inthe development of atherosclerosis, and is characterized by an increasedpermeability of endothelium, tendency to vasospasm and thrombosis.

Apolipoprotein A5, a minor plasma apolipoprotein, has been docu-mented to play a major role in triglyceride metabolism by the enhance-ment of very low density lipoprotein (VLDL) lipolysis and clearance, andinhibition of VLDL production [5]. The Apo A5 gene was identified 30 kbupstream of the well-characterized Apo A1/C3/A4/A5 gene cluster onchromosome 11q23, and comprises 3 exons encoding 366 amino acids.The −1131C Apo A5 allele of polymorphism in the gene promoter wasidentified as a susceptibility variant for development of MetS in recentstudies in different populations [6–11], although the results are not fullyconsistent [12]. Nevertheless, the association between Apo A5 T-1131C(rs662799) polymorphism and hemostatic markers, and their potentialrelationship to MetS has not been evaluated so far.

12 D. Novotny et al. / Clinica Chimica Acta 436 (2014) 11–17

Apolipoprotein E is a structural component of triglyceride-rich lipo-proteins serving as a high-affinity ligand for low density lipoprotein(LDL) receptor and related proteins. It plays also an important role inthe catabolism of remnant lipoprotein particles [13]. Apo E4 (Cys112Arg,rs429358) and Apo E2 (Arg158Cys, rs7412) variants differ from commonApo E3 isoform by a single amino acid substitution and the variants varyon both structural and functional levels. In recent studies in different pop-ulations, Apo E4 variant was associated with both carotid and coronaryatherosclerosis [14–18], carotid plaque formation [17], cerebral infarction[18], and other atherosclerosis-linked syndromes, although the resultsare conflicting in some works (e.g. [19]). Several studies also evaluatedan association between ApoE polymorphism and metabolic syndrome[20–23]. Nevertheless, very little is known about the relationship ofhemostatic markers to Apo E polymorphism in asymptomatic dyslipid-emic patients.

Therefore, our study aimed to evaluate the relationship of T-1131CApo A5 (rs662799) variants, and variants of Apo E common polymor-phism (rs429358, rs7412) to fibrinogen, von Willebrand factor, tissueplasminogen activator, and plasminogen activator–inhibitor 1 levels indyslipidemic patients, subsequently divided according to the presence/absence of MetS. Together with these parameters, lipid and lipoproteinanalytes as well as markers of insulin resistance were determined.

2. Materials and methods

2.1. Study design and subjects

The study was performed with asymptomatic dyslipidemic subjects(individuals without a history of clinically manifest atherosclerosis —coronary artery disease, heart failure, cerebrovascular ischemic diseaseand peripheral vascular disease, with altered plasma lipids, i.e.with total cholesterol (TC) ≥5.0 mmol/l and/or triglycerides (TG)≥1.5 mmol/l). They had been examined for the first time in theLipid Centre of the 3rd Department of Internal Medicine, UniversityHospital Olomouc, Czech Republic, during the period from January2006 to March 2011. All subjects were tested for the signs of secondaryhyperlipidemia: diabetes mellitus, hypothyroidism, renal or hepaticdiseases and nephrotic syndrome. Other exclusion criteria were asfollows: history of clinically manifested atherosclerosis presented bycoronary artery disease, cerebrovascular disease and peripheral arterialdisease, hypolipidemic therapy in the previous 8 weeks, hormonetherapy and the clinical presence of acute infections. The subjects withE24 genotype of Apo E polymorphism were also excluded because ofthe opposite effects of E2 and E4 alleles on lipid levels. All individualsfilled out a questionnaire on their previous medical history, especiallycardiovascular status, medication and smoking habits. Body massindex and systolic and diastolic blood pressures (SBP, DBP) were alsodetermined. The study was reviewed and approved by the EthicsCommittee of the Medical Faculty and University Hospital Olomoucand informed consent was obtained from all participants.

Individuals who met the criteria mentioned above (n = 590, 288males, 302 females) were subsequently divided into two groups: patientswith the presence ofmetabolic syndrome (MetS+, n=146, 77males, 69females), and individuals with the absence of metabolic syndrome(MetS−, n = 444, 211 males, 233 females). NCEP ATPIII Panel 2001criteria were used for identification of MetS and the presence of at leastthree of the factors was sufficient for the diagnosis of MetS [24].

2.2. Laboratory analysis

Venous blood samples were drawn in the morning after a 12-h fast.After centrifugation, the serum was used for other analyses. For theassessment of hemostatic markers, venous blood was collected in 3.8%sodium citrate tubes and plasma was obtained after centrifugation.

Routine serum biochemical parameters were analyzed on aModularSWA system (Roche, Basel, Switzerland) in the same day of blood

collection. Concentrations of other special analytes were measuredin the sample aliquots stored at −80 (−20) °C, no longer than6 months — see below in the text.

Total cholesterol, TG andHDLc levelswere determined enzymaticallyon aModular SWA system (Roche, Basel, Switzerland). Low density lipo-protein cholesterol (LDLc) levels were calculated using the Friedewaldformula (for TG less than 4.5 mmol/l). Other calculated parameterswere as follows: nonHDL-cholesterol (nonHDLc = TC− HDLc), athero-genic index of plasma (AI) (logTG/HDLc), andhomeostasismodel assess-ment (HOMA-R= glucose × insulin/22.5). Concentrations of Apo B andApo A1 were determined immunoturbidimetrically using Tina-Quant ApoB and ApoA-1 kits (Roche, Basel, Switzerland). Lipopro-tein (a) [Lp(a)] was measured immunoturbidimetrically using Tina-Quant lipoprotein(a) TQ kit (Roche, Basel, Switzerland). C-reactiveprotein (CRP) was assessed by an ultrasensitive immunoturbidimetricmethod using the kit Tina-Quant (Roche, Basel, Switzerland). Glu-cose was determined using the GOD-PAP method (Roche, Basel,Switzerland). All tests were measured from fresh sera in the sameday of blood collection.

Insulin was determined by the commercially available kit(Immunotech, Marseille, France) using specific antibodies by the IRMAmethod. C-peptide and proinsulin (PINS) were determined using thefollowing kits: C-peptide (Immunotech,Marseille, France), and proinsu-lin (DRG Instruments GmbH,Marburg, Germany), by the IRMAmethod,and RIA method, respectively. The sample aliquots were storedat −20 °C, no longer than 6 months.

The following hemostatic markers were examined from humanplasma stored at −20 °C: fibrinogen (function coagulation method byClauss, Technoclone, Vienna, Austria), von Willebrand factor (immuno-turbidimetric assay, Instrumentation Laboratory Spa, Milan, Italy), plas-minogen activator inhibitor-1, and tissue plasminogen activator (ELISA,Technoclone, Vienna, Austria).

2.3. Genotyping

DNA was extracted from the peripheral leukocytes of all subjectsusing a standard commercial kit (QIAamp DNA blood mini-kit, Qiagen,Germany). After isolation, the extractswere stored at−20 °C, no longerthan 3 months. Apo A5 genotypes were determined by a melting curveanalysis after a real time PCRmethod adapted from the work of Franceset al. [25]. Genotyping of Apo E alleles was performed using a commer-cially available Apo E LightMix kit (Roche, Basel, Switzerland) by amelt-ing curve analysis after real time PCR. The 228 bp fragments wereamplified with specific primers and PCR amplicons then analyzedusing SimpleProbe and hybridization probes.

2.4. Statistical analysis

All values of quantitative parameters are expressed asmeans± stan-dard deviation (SD) and parameters with skewed distribution alsoexpressed as medians. The Kolmogorov–Smirnov test was used tocheck for normal distribution. Variables with skewed distribution [CRP,TG, AI, Lp(a), tPA, C-peptide] were log transformed in order to normalizetheir distribution before statistical analysis. Differences in variablesbetween individual groupswere analyzedwith ANOVA, after adjustmentfor age and sex. Statistical analysis was performed by SPSS forWindows,version 12.0 (SPSS Inc., Chicago, IL, USA). Probability values of p b 0.05were considered as statistically significant.

The genotype and allele frequencies for the Apo A5 and Apo E genesamong all subjects together with MetS+ and MetS− groups weredetermined by gene countingmethod, and then evaluated by performingPearson Chi-square statistical analysis to evaluate whether followedpolymorphisms were consistent with Hardy–Weinberg equilibriumexpectation.

13D. Novotny et al. / Clinica Chimica Acta 436 (2014) 11–17

3. Results

The basic clinical and laboratory characteristics of all subjects, MetS−and MetS+ groups are summarized in Table 1. MetS+ subjects wereolder in comparison to MetS− individuals, with higher systolic and dia-stolic blood pressures, and unfavorable anthropometrical parameters,such aswaist circumference and bodymass index. Aside from the expect-ed altered lipid and lipoprotein profiles (elevated TC, TG, nonHDLc, AI,Apo B, and decreased levels of HDLc and Apo A1), the individuals withMetS had strongly significantly higher levels of CRP, tPA, PAI-1, glucose,insulin, HOMA-R, C-peptide, and proinsulin (all p b 0.001).

In Table 2, differences in hemostatic markers in all groups accordingto Apo A5 SNP T-1131C (rs662799) variants are introduced. We have

Table 1Characteristics of all individuals and subjects without and with MetS.

All individuals n = 590

Male/female 288/302Age years 45.2 ± 14.9

(33–57)SBP mm Hg 129.2 ± 16.0

(120–140)DBP mm Hg 79.7 ± 9.3

(75–85)Waist cm 86.8 ± 12.8

(77–96)BMI kg/m2 25.76 ± 4.01

(23.0–28.1)CRP mg/l 2.5 ± 3.5 [1.4]

(0.7–3.0)TC mmol/l 6.44 ± 2.06

(5.17–7.39)TG mmol/l 2.49 ± 2.85 [1.71]

(1.13–2.60)AI 0.1029 ± 0.3728 [0.0639]

(−0.1622–0.3126)nonHDL mmol/l 4.92 ± 1.91

(3.69–5.84)HDLc mmol/l 1.49 ± 0.44

(1.19–1.73)LDLc mmol/l 3.90 ± 1.62

(2.86–4.70)Apo A1 g/l 1.58 ± 0.33

(1.38–1.75)Apo B g/l 1.16 ± 0.34

(0.93–1.37)Lp(a) g/l 0.337 ± 0.527 [0.168]

(0.065–0.557)Fibrinogen g/l 3.39 ± 3.00

(2.76–3.70)vWF % 133 ± 51

(99–159)tPA ng/ml 4.00 ± 5.77 [2.3]

(1.50–4.80)PAI-1 ng/ml 73 ± 39

(43–97)Glucose mmol/l 5.10 ± 0.95

(4.57–5.40)Insulin mIU/l 8.5 ± 5.6

(5.2–10.6)HOMA-R 1.981 ± 1.754

(1.097–2.380)C-peptide mg/l 6.0 ± 64 [2.1]

(1.5–2.9)PINS mIU/l 13.8 ± 10.4

(8.3–15.7)

Data are expressed as means ± standard deviations, in parameters with skewed distribution aDifferences in variables between groups were analyzed with ANOVA after adjustment for agetransformed to normalize their distribution before statistical analyses. Significant differences beSBP, systolic blood pressure; DBP, diastolic blood pressure; BMI, body mass index; CRP, C-rcholesterol; LDLc, low density lipoprotein cholesterol; nonHDL = TC − HDLc; AI, atherogenWillebrand factor; tPA, tissue plasminogen activator; PAI-1, plasminogen activator inhibitor-1;

observed the difference between tPA levels in all individuals andMetS− groups, where −1131C minor variant was associated with ele-vated tPA in comparison to TT genotype (p b 0.001). But this was notthe case of MetS+ group, in which no significant variance was seenin tPA levels. However, in this group a weak difference was observedin PAI-1 levels, with decreased PAI-1 in −1131C minor variant carriers(p b 0.05). No significant changes in concentration of other hemostaticmarkers (fibrinogen, vWF) were seen in all three groups.

According to Apo E polymorphism (rs429358, rs7412), statisticallynon-significant increased tPA levels in all individuals, MetS− andMetS+groupswere observed in connectionwithminor Apo E2 variant,compared to “neutral” Apo E33 genotype (Table 3). Significantly higherPAI-1 concentrations in MetS+ group (see Table 1) were accompanied

MetS− n = 444 Met S+ n = 146

211/233 77/6942.3 ± 15.0 53.7 ± 10.8***(31–55) (47–61)126.0 ± 14.7 138.9 ± 16.2***(120–134) (130–145)78.1 ± 9.0 85 ± 8.0***(70–80) (80–90)83.0 ± 11.3 97.4 ± 10.6***(74–91) (91–105)24.49 ± 3.33 29.41 ± 3.56***(22.3–26.6) (27.1–31.6)2.2 ± 3.3 [1.2] 3.3 ± 3.8 [2.2]***(0.6–2.5) (1.2–4.0)6.22 ± 1.75 6.98 ± 2.35***(4.99–7.21) (5.77–7.61)1.88 ± 1.91 [1.40] 4.01 ± 3.79 [2.88]***(1.01–2.05) (2.05–4.48)−0.0056 ± 0.3247 [−0.0313] 0.4254 ± 0.3201 [0.3782]***(−0.2265–0.1586) (0.1950–0.6586)4.65 ± 1.76 5.72 ± 2.13***(3.47–5.61) (4.63–6.36)1.57 ± 0.42 1.27 ± 0.44***(1.26–1.79) (0.96–1.46)3.81 ± 1.43 4.14 ± 2.03*(2.68–4.71) (3.30–4.67)1.63 ± 0.33 1.47 ± 0.29***(1.41–1.81) (1.25–1.63)1.12 ± 0.34 1.28 ± 0.30***(0.87–1.34) (1.09–1.45)0.398 ± 0.468 [0.185] 0.316 ± 0.676 [0.129](0.077–0.624) (0.057–0.350)3.37 ± 3.45 3.43 ± 0.71(2.73–3.64) (2.90–3.87)130 ± 48 140 ± 57(99–155) (100–166)3.60 ± 5.97 [2.1] 5.28 ± 4.99 [3.0]***(1.30–3.83) (2.00–7.10)67 ± 35 89 ± 44***(40–90) (58–108)4.86 ± 0.57 5.81 ± 1.36***(4.50–5.20) (5.00–6.10)7.3 ± 4.3 12.1 ± 7.5***(4.5–8.9) (7.2–15.2)1.587 ± 1.003 3.167 ± 2.736***(0.940–1.972) (1.809–3.860)3.8 ± 36 [1.9] 12.6 ± 111.7 [3.2]***(1.4–2.4) (2.4–4.1)11.6 ± 6.6 20.8 ± 15.9***(7.8–13.3) (11.3–24.5)

lso as medians [in brackets]. Values of 25 and 75 percentiles are expressed in parentheses.and sex. Variables with skewed distribution [CRP, TG, AI, Lp(a), tPA, C-peptide] were logtween MetS− and MetS+ groups: *p b 0.05, **p b 0.01 and ***p b 0.001. Abbreviations:eactive protein; TC, total cholesterol; TG, triglycerides; HDLc, high density lipoproteinic index of plasma (logTG/HDLc); Apo, apolipoprotein; Lp(a), lipoprotein (a); vWF, vonHOMA-R, homeostasis model assessment; and PINS, proinsulin.

Table 2Differences in hemostatic markers in all individuals, MetS− and MetS+ groups according to Apo A5 T-1131C (rs662799) variants.

Apo A5 (rs662799) All individuals n = 590 MetS− n = 444 MetS+ n = 146

−1131T (TT)n = 485

−1131C (TC + CC)n = 105

−1131T (TT)n = 368

−1131C (TC + CC)n = 76

−1131T (TT)n = 120

−1131C (TC + CC)n = 26

Fibrinogen g/l 3.42 ± 3.29 3.25 ± 0.74 3.43 ± 3.78 3.11 ± 0.65 3.39 ± 0.66 3.64 ± 0.89vWF % 133 ± 52 132 ± 46 130 ± 49 130 ± 42 140 ± 58 141 ± 57tPA ng/ml 3.67 ± 3.95 (2.20) 5.61 ± 10.65 (3.00)*** 3.14 ± 3.36 (2.00) 5.75 ± 12.12 (2.70)*** 5.32 ± 5.09 (3.00) 5.38 ± 4.73 (3.45)PAI-1 ng/ml 73 ± 39 74 ± 38 66 ± 34 73 ± 40 93 ± 45 76 ± 31*

Data are expressed as means ± standard deviations, in parameters with skewed distribution also as medians (in parentheses). Differences in variables between groups were analyzed withANOVA after adjustment for age and sex. Variables with skewed distribution (tPA) were log transformed to normalize their distribution before statistical analyses. Significant differences invariables between Apo A5 genotypes in all individuals, MetS− andMetS+ groups: *p b 0.05, **p b 0.01 and ***p b 0.001. Abbreviations: vWF, vonWillebrand factor; tPA, tissue plasminogenactivator; and PAI-1, plasminogen activator inhibitor-1.

14 D. Novotny et al. / Clinica Chimica Acta 436 (2014) 11–17

bynon-significant elevation of PAI-1 in both Apo E2 andApo E4 carriers.No significant differences offibrinogen and vWF levelswere observed inall three groups.

The relationships between Apo A5/Apo E common variants andhemostatic markers are presented in Table 4. The analysis was performedonly in the whole group of all dyslipidemic patients because of low num-ber of participants in Apo A5/Apo E subgroups. The results indicate thatthe presence of−1131C minor Apo A5 allele has always been associatedwith higher levels of tPA in comparison with T allele, regardless of Apo Egenotype. Also the presence of minor Apo E2 allele led to elevated tPAin both T and C carriers of T-1131C Apo A5 polymorphism. In addi-tion, the−1131C/E2 variant seems to be associated with the highesttPA levels.

The distribution of combination of Apo A5 and Apo E variants andcombined Apo A5/Apo E variants in the study groups is summarizedin Table 5. Genotype frequencies of polymorphisms at both geneswere shown to be in Hardy–Weinberg equilibrium for each group. Asshown in Table 5, no significant differences were observed betweenall groups of dyslipidemic subjects in both Apo A5 and Apo E genotypefrequencies, and also in common Apo A5/Apo E variant frequencies.

4. Discussion

The present study confirms the well-established risk profile ofsubjects with MetS, although clinically asymptomatic. In addition, twohemostatic markers, tPA and PAI-1, were significantly elevated, whichmight reflect enhanced procoagulation state aswell as impaired fibrino-lysis and endothelial dysfunction in MetS patients.

In prospective epidemiological studies, elevated levels of fibrinogenhave consistently been shown to be a strong independent cardiovascu-lar risk factor [26,27]. Nevertheless, the relationship between fibrinogenand features of the insulin resistance syndrome is weaker than that ofother hemostatic factors [28]. In accordance with this finding, wefound no differences in fibrinogen concentrations between MetS−and MetS+ subjects.

vWF levels independently predicted intimamedia thickness, a mark-er of subclinical atherosclerosis, in asymptomatic dyslipidemic patientsin our previous work [29]. In the present study, only a non-significantelevation of vWF levels in subjectswithMetSmay reflect proatherogenicconditions in both MetS− and MetS+ groups.

4.1. Tissue plasminogen activator and plasminogen activator-inhibitor 1

Clinical studies have identified tPA as a strong predictor of futurecardiac events [28]. Previouslywe demonstrated higher tPA in hyperlip-idemic subjects and asymptomatic members of families with familialcombined hyperlipidemia [30], and dyslipidemic subjects with elevatedtriglycerides and Apo B [31]. The association between tPA levels andmetabolic syndromemay reflect the elevation in PAI-1 levels that occursin this condition, given that the majority of tPA circulates in the plasmain complex with PAI-1. However, as increased tPA occurs in connection

with endothelial dysfunction, elevated levels may reflect the presenceof underlying endothelial damage [32,33]. Thus, elevated tPA levelsin our study may suggest the presence of a proatherogenic conditionenhanced in individuals with features of metabolic syndrome, withoutthe respect to PAI-1 levels.

PAI-1 is involved in the regulation of fibrinolysis inhibiting the princi-pal activator of plasminogen, tPA [28]. Elevated PAI-1 levels were anindependent risk factor for the development of type 2 diabetes mellitusin healthy subjects in the IRAS Study [34], and may be an early riskmarker for the development of the metabolic syndrome. Clinical studiessuggest that visceral adipose tissue is the main determinant of elevatedPAI-1 in humans [35]. Therefore, increased PAI-1 gene expression inobesity may explain elevated levels of PAI-1 in MetS+ subjects.

4.2. Apo A5 (rs662799) polymorphism

In the present study, Apo A5 variant frequencies were similar to thosepreviously published in the Caucasian population in all subgroups. Theevaluation of hemostatic factors according to Apo A5 SNP T-1131C vari-ants revealed the significant difference between tPA levels in the wholeand MetS− groups (but not in MetS+), where CC and CT genotypeswere associated with elevated tPA compared to TT variant. The presenceof C allele did not affect the levels of other hemostatic parameters in com-parison to TT genotype in all groups except for MetS+ subjects, whereT allele was weakly related with higher PAI-1 levels.

Apolipoprotein A5 gene variants have been shown to be associatedwith triglyceride levels. Numerous studies have confirmed the associationof naturally occurring variants of the Apo A5 gene with increased TG, in-cluding−1131C variant [36–38,5]. Several mechanisms of Apo A5 actionon the protein level have been proposed describing an effect on plasmaTG removal: stimulation of lipoprotein lipase (LPL)-mediated triglyceridehydrolysis, acceleration of hepatic uptake of triglyceride-rich lipoproteinsand their remnants, and intracellular effect on hepatic VLDL productionand/or secretion [39]. In the study of Kim et al., the presence of ApoA5 −1131C rare variant led to reduced Apo A5 plasma concentrationswith concomitantly restricted LPL activation, resulting in higher serumTG levels, smaller LDL particle size, and lower HDL cholesterol levels[40]. Alternatively, changes in the expression and/or activity of nuclearreceptors regulating Apo A5 expression might also contribute to de-creased plasma Apo A5 levels in patients with metabolic diseases [39].

Higher plasma concentrations of TG in −1131C allele carriers mayitself reflect an increased risk of endothelial dysfunction, as shown insome experimental and clinical studies [41,42]. In our study, the multi-ple regression analysis with tPA as a dependent variable revealed closeassociation with TG levels (b = 0.411, p = 0.0001), and C allele wasassociated with elevated tPA compared to TT variant only in MetS−subjects. These relationships have not been confirmed inMetS+ patients(data not published). However, we speculate about the low number ofparticipants in the MetS+ group and mainly C allele carriers, in whichhigher levels of tPA are indicated in comparison to TT homozygotes. Itis possible that the increase of the number of MetS+ subjects would

Table 3Differences in hemostatic markers in all individuals, MetS− and MetS+ groups according to Apo E (rs429358, rs7412) variants.

Apo E (rs429358,rs7412)

All individualsn = 590

MetS− n = 444 MetS+ n = 146

E2 (E22 + E23)n = 85

E3 (E33)n = 335

E4 (E34 + E44)n = 170

E2 (E22 + E23)n = 66

E3 (E33)n = 248

E4 (E34 + E44)n = 130

E2 (E22 + E23)n = 19

E3 (E33)n = 87

E4 (E34 + E44)n = 40

Fibrinogen g/l 3.22 ± 0.66 3.48 ± 3.95 3.26 ± 0.69 3.14 ± 0.56 3.50 ± 4.60 3.21 ± 0.71 3.45 ± 0.89 3.41 ± 0.70 3.44 ± 0.63vWF % 134 ± 57 134 ± 53 130 ± 43 137 ± 59 130 ± 48 127 ± 41 126 ± 49 144 ± 62 139 ± 50tPA ng/ml 5.06 ± 5.58

(3.00)3.82 ± 6.47(2.30)

3.92 ± 4.29(2.20)

4.39 ± 5.18(2.85)

3.46 ± 7.09(2.00)

3.54 ± 3.73(2.00)

7.25 ± 6.39(5.50)

4.91 ± 4.25(3.00)

5.15 ± 5.63(3.10)

PAI-1 ng/ml 78 ± 42 69 ± 37 78 ± 40 73 ± 36 65 ± 34 71 ± 38 93 ± 54 83 ± 42 102 ± 39

Data are expressed as means ± standard deviations, in parameters with skewed distribution also as medians (in parentheses). Differences in variables between groups were analyzed withANOVA after adjustment for age and sex. Variables with skewed distribution (tPA) were log transformed to normalize their distribution before statistical analyses. Significant differences invariables between Apo A5 genotypes in all individuals, MetS− andMetS+ groups: *p b 0.05, **p b 0.01 and ***p b 0.001. Abbreviations: vWF, vonWillebrand factor; tPA, tissue plasminogenactivator; and PAI-1, plasminogen activator inhibitor-1.

15D. Novotny et al. / Clinica Chimica Acta 436 (2014) 11–17

emphasize the effect of C allele more on both tPA levels and the relation-ship between tPA and TG. Nevertheless, it seems probable that thepresence of Apo A5 −1131C variant is associated with both higher TGand tPA levels in dyslipidemic patients, without any respect to theabsence/presence of the metabolic syndrome. Further studies with thedifferent populations are needed to confirm if both parameters aredependent on each other in this relationship. In any case, to the best ofour knowledge, our study is the first to describe the association of Callele of Apo A5 (rs662799) polymorphism with elevated levels of tPAcompared to wildtype TT variant in dyslipidemic subjects.

4.3. Apo E (rs429358, rs7412) polymorphisms

In our study we demonstrated that Apo E genotype distribution inMetS+ is similar to that observed in the whole and MetS− groups. Incomparison to the general Czech population [43], the frequency of E4 var-iant was especially higher in all three dyslipidemic subgroups whichcould reflect an increased risk of proatherogenic conditions in all the sub-jects. However, aside from studies mentioned above, recent meta-analysis of 14.015 stroke cases revealed the association of Apo E2 variantswith ischemic strokewith odds ratios 1.09 for E22, and 1.05 for E24 geno-types [44]. Also the case–control study of Nagato et al. [45] demonstratedan association between ApoE 2 variant and the deep venous thromboticevents in women. The Rotterdam study suggested that E4 variant ofApo E is not an important risk factor for carotid artery atherosclerosis[19], and any protective effect of E23 genotype is limited. In our study,no significant differences in levels of all monitored hemostatic markerswere observed between Apo E genotypes in all three groups. Neverthe-less, the presence of Apo E2 variant was associated with non-significantelevation of tPA levels in all groups compared to “neutral” Apo E33genotype suggesting the possible role of Apo E2 variant as a potentialprocoagulant parameter risk.

Table 4Differences in hemostatic markers of all individuals (n = 590) according to variants of Apo A5lipidemic patients in comparison to variant −1131T/E3.

Apo A5/ApoE variants

−1131T/E3(TT/E33)n = 286

−1131T/E2(TT/E22 + E23)n = 65

−1131T/E4(TT/E34 + E44)n = 135

Fibrinogen g/l 3.51 ± 4.30 3.30 ± 0.60 3.25 ± 0.70vWF % 133 ± 54 137 ± 61 132 ± 43tPA ng/ml 3.48 ± 3.59 (2.20) 4.81 ± 5.41 (2.85)* 3.56 ± 3.86 (2.15)PAI-1 ng/ml 70 ± 37 77 ± 43 78 ± 41

Data are expressed as means ± standard deviations, in parameters with skewed distributionwith ANOVA after adjustment for age and sex. Variables with skewed distribution (tPA) were loences between Apo A5/Apo E variants: *p b 0.05, **p b 0.01 and ***p b 0.001. Abbreviations: vtivator inhibitor-1.

4.4. Apo A5/Apo E common variants

The majority of genetic polymorphisms exert their effects in accor-dance with other polymorphisms as haplogroups. Therefore, we alsoevaluated a relationship between Apo A5/Apo E combined variants tothe levels of hemostatic factors. To the best of our knowledge, there isno information in recent literature describing relations mentionedabove, and sowe have no data for comparison. Our observations confirmthe findings of the above-mentioned association of theminor alleleswithincreased levels of tPA. In addition,−1131C/E2 variant was strongly as-sociated with the highest tPA levels emphasizing their effects, whichcould represent a new possible risk parameter of prothrombotic condi-tions in dyslipidemic individuals. Themechanism of potential interactionbetween Apo E and Apo 5 variants in genetic determination of tPA levelsneeds to be analyzed in further studies. To obtain relevant conclusions,the analysis should be performed with a larger number of participants.

4.5. Limitation of the study

The present study is limited by the relatively low number of subjectsin Apo A5/Apo E variant subgroups which does not allow the perfor-mance of statistical evaluation according to the presence/absence ofMetS, and reduces the power of significance. For the same reason wedid not evaluate female and male subjects separately in all observedsubgroups, although we are aware of the different effects especially ofApo E polymorphism on both laboratory results and clinical outcomes.

5. Conclusion

Our study has first evaluated the influence of apolipoprotein A5 andapolipoprotein E genetic variants on the levels of hemostatic markers inasymptomatic dyslipidemic patients, both without and with features of

SNP T-1131C (rs662799) and Apo E common polymorphisms (rs7412, rs429358) in dys-

−1131C/E2(TC + CC/E22 + E23)n = 19

−1131C/E3(TC + CC/E33)n = 54

−1131C/E4(TC + CC/E34 + E44)n = 31

2.98 ± 0.78 3.27 ± 0.73 3.31 ± 0.70129 ± 41 141 ± 49 119 ± 446.19 ± 6.28 (4.55)** 5.52 ± 13.74 (2.70)* 5.61 ± 5.68 (2.88)**80 ± 39 69 ± 37 80 ± 37

also as medians (in parentheses). Differences in variables between groups were analyzedg transformed to normalize their distribution before statistical analyses. Significant differ-WF, von Willebrand factor; tPA, tissue plasminogen activator; and PAI-1, plasminogen ac-

Table 5The distribution of Apo A5/Apo E variants in all individuals, MetS− and MetS+ groups.

All individuals n = 590 MetS− n = 444 MetS+ n = 146

Apo A5 (rs662799) variants−1131T (TT) 488 (83%) 368 (83%) 120 (82%)−1131C (TC + CC) 102 (17%) 76 (17%) 26 (18%)

Apo E (rs429358, rs7412) variantsE2 (E22 + E23) 85 (14%) 66 (15%) 19 (13%)E3 (E33) 335 (57%) 248 (56%) 87 (59%)E4 (E34 + E44) 170 (29%) 130 (29%) 40 (28%)

Apo A5/Apo E variants−1131T/E2 (TT/E22 + E23) 65 (11%) 52 (12%) 13 (9%)−1131T/E3 (TT/E33) 286 (49%) 216 (49%) 70 (48%)−1131T/E4 (TT/E34 + E44) 135 (23%) 99 (22%) 36 (25%)−1131C/E2 (TC + CC/E22 + E23) 19 (3%) 13 (3%) 6 (4%)−1131C/E3 (TC + CC/E33) 54 (9%) 36 (8%) 18 (12%)−1131C/E4 (TC + CC/E34 + E44) 31 (5%) 28 (6%) 3 (2%)

Apo A5 (rs662799): Pearson Chi-square 0.704, p = 0.703 (MetS− vs. MetS+). Apo E (rs429358, rs7412): Pearson Chi-square 0.554, p = 0.758 (MetS− vs. MetS+). Apo A5/Apo E var-iants: Pearson Chi-square 6.914, p = 0.227 (MetS− vs. MetS+). Non-significant difference in variant frequencies was observed between MetS− and MetS+ groups.

16 D. Novotny et al. / Clinica Chimica Acta 436 (2014) 11–17

metabolic syndrome.Wehave demonstrated especially the strong associ-ation of higher tPA levelswith theminor−1131CApoA5 variant, but notin patients withmetabolic syndrome. The presence of Apo E2 variant wasassociated with non-significant elevation of tPA levels in all groups.Newly we have also described the analysis of common Apo A5/Apo E ge-netic variants,where the−1131C/E2 combination seems to be associatedwith high-risk tPA levels. We conclude that this variant could represent anew possible risk parameter of prothrombotic conditions in dyslipidemicpatients, regardless of the presence/absence of the metabolic syndrome.This observation needs to be clarified in further research.

Conflict of interest statement

None.

Acknowledgment

This study is supported by the Institutional Support of Ministry ofHealth, Czech Republic, Nr.1 RVO-FNOL2013.

References

[1] Samad F, Ruf W. Inflammation, obesity, and thrombosis. Blood 2013;122:3415–22.[2] Eckel RH, Grundy SM, Zimment PZ. The metabolic syndrome. Lancet

2005;365:1415–28.[3] Alessi MC, Juhan-Vague I. Metabolic syndrome, haemostasis and thrombosis.

Thromb Haemost 2008;99(6):995–1000. http://dx.doi.org/10.1160/TH07-11-0682.[4] Kostapanos M, Florentin M, Elisaf MS, Mikhailidis DP. Hemostatic factors and the

metabolic syndrome. Curr Vasc Pharmacol 2013;11:880–905. http://dx.doi.org/10.2174/15701611113116660171.

[5] Sharma V, Ryan RO, Forte TM. Apolipoprotein A-V dependent modulation of plasmatriacylglycerol: a puzzlement. Biochim Biophys Acta 1821;2012:795–9.

[6] Yamada Y, Kato K, Hibino T, et al. Prediction of genetic risk for metabolic syndrome.Atherosclerosis 2007;191:298–304.

[7] Grallert H, Sedlmeier EM, Huth C, et al. APOA5 variants and metabolic syndrome inCaucasians. J Lipid Res 2007;48:2614–21.

[8] Maasz A, Kisfali P, Horvatovich K, et al. Apolipoprotein A5 T-1131C variant confersrisk for metabolic syndrome. Pathol Oncol Res 2007;13:243–7.

[9] Ong KL, Jiang CQ, Liu B, et al. Association of a genetic variant in the apolipoprotein A5gene with the metabolic syndrome in Chinese. Clin Endocrinol 2011;74:206–13.

[10] Garelnabi M, Lor K, Jin J, Chai F, Santanam N. The paradox of ApoA5 modulation oftriglycerides: evidence from clinical and basic research. Clin Biochem 2013;46:12–9.

[11] Zaki M, Amr K. Apolipoprotein A5 T-1131C variant and risk for metabolic syndrome inobese adolescents. Gene 2014;634:44–7. http://dx.doi.org/10.1016/j.gene.2013.10.032.

[12] Ch Xu, Bai R, Zhang D, et al. Effects of APOA5–1131T/C (rs662799) on fastingplasma lipids and risk of metabolic syndrome: evidence from a case–controlstudy in China and a meta-analysis. PLoS ONE 2013;8:e56216. http://dx.doi.org/10.1371/journal.pone.0056216.

[13] Wilson PWF, Schaefer EJ, LarsonMG, Ordovas JM. Apolipoprotein E alleles and risk ofcoronary disease: a meta analysis. Arterioscler Thromb Vasc Biol 1996;16:1250–5.

[14] Graner M, Kahri J, Varpula M, et al. Apolipoprotein E polymorphism is associatedwith both carotid and coronary atherosclerosis in patients with coronary arterydisease. Nutr Metab Cardiovasc Dis 2008;18:271–7.

[15] Debette S, Lambert JC, Gariépy J, et al. New insight into the association of apolipopro-tein E genetic variants with carotid plaques and intima-media thickness. Stroke2006;37:2917–23.

[16] Atabek ME, Ozkul Y, Eklioğlu BS, Kurtoğlu S, Baykara M. Association between apolipo-protein E polymorphism and subclinic atherosclerosis in patients with type 1 diabetesmellitus. Clin Res Pediatr Endocrinol 2012;4(1):8–13. http://dx.doi.org/10.4274/jcrpe.521.

[17] Beilby JP, Hunt CCJ, Palmer LJ, et al. Apolipoprotein E gene polymorphisms are asso-ciated with carotid plaque formation but not with intima-media wall thickening.Stroke 2003;34:869–74.

[18] Wang QY, Wang WJ, Wu L, Liu L, Han LZ. Meta-analysis of APOE e2/e3/e4 polymor-phism and cerebral infarction. J Neural Transm 2013;120:1479–89. http://dx.doi.org/10.1007/s00702-013-1019-8.

[19] Slooter AJC, Bots ML, Havekes LM, et al. Apolipoprotein E and carotid arteryatherosclerosis. The Rotterdam Study. Stroke 2001;32:1947–52.

[20] Sima A, Iordan A, Stancu C. Apolipoprotein E polymorphism — a risk factor formetabolic syndrome. Clin Chem Lab Med 2007;45:1149–53.

[21] Lee KS, Jang Y, Chung YK, Chung JH, Oh BH, Hong CH. Relationship between thediagnostic components of metabolic syndrome (MS) and cognition by ApoEgenotype in the elderly. Arch Gerontol Geriatr 2010;50:69–72.

[22] Ferreira DC, Costa TF, Aguiar SFL, et al. Association of apoliprotein E polymor-phisms and metabolic syndrome in subjects with extreme obesity. Clin ChimActa 2011;412:1559–62.

[23] Luptakova L, Sivakova D, Cvicelova M, et al. Power of biomarkers and theirrelative contributions to metabolic syndrome in Slovak adult women. AnnHum Biol 2013;40:132–8.

[24] Executive summary of the third report of the National Cholesterol EducationProgram (NCEP) Expert Panel on Detection, Evaluation, And Treatmentof High Blood Cholesterol In Adults (Adult Treatment Panel III). JAMA2001;285:2486–97.

[25] Frances F, Corella D, Sorli JV, Guillen M, Gonzalez JI, Portoles O. Validating a rapidmethod for detecting common polymorphisms in the APOA5 gene by meltingcurve analysis using LightTyper. Clin Chem 2005;51:1279–82.

[26] Koenig W. Fibrin(ogen) in cardiovascular disease: an update. Thromb Haemost2003;89:601–9.

[27] Ernst E, Resch K. Fibrinogen as a cardiovascular risk factor: a meta-analysis andreview of the literature. Ann Intern Med 1993;118:956–63.

[28] Grant PJ. Diabetes mellitus as a prothrombotic condition. J Intern Med2007;262:157–72.

[29] Karasek D, Vaverkova H, Halenka M, et al. Prothrombotic markers in asymptomaticdyslipidemic subjects. J Thromb Thrombolysis 2011;31:27–36. http://dx.doi.org/10.1007/s11239-010-0474-4.

[30] Karásek D, Vaverková H, Halenka M, Slavík L, Novotný D. Endothelial haemostaticmarkers in members of families with familial combined hyperlipidemia. ThrombRes 2009;123:466–75.

[31] Vaverkova H, Karasek D, Novotny D, et al. Positive association of adiponectinwith soluble thrombomodulin, von Willebrand factor and soluble VCAM-1 indyslipidemic subjects. Clin Biochem 2013;46:766–71. http://dx.doi.org/10.1016/j.clinbiochem.2013.02.014.

[32] Steins M, Padro T, Li C, Mesters RM, Ostermann H, Hammel D. Overexpression oftissue-type plasminogen activator in atherosclerotic human coronary arteries.Atherosclerosis 1999;145:173–80.

[33] Lowe GD, Danesh J, Lewington S,Walker M, Lennon L, Thomson A. Tissue plasminogenactivator antigen and coronary heart disease. Prospective study and meta-analysis. EurHeart J 2004;25:252–9.

17D. Novotny et al. / Clinica Chimica Acta 436 (2014) 11–17

[34] Festa A, D'Agostino R, Tracy R, Haffner S. Elevated levels of acute-phase proteinsand plasminogen activator inhibitor-1 predict the development of type 2 diabetes.Diabetes 2002;51:1131–7.

[35] Mertens I, Van der PlankenM, Corthouts B, et al. Visceral fat is a determinant of PAI-1 activity in diabetic and non-diabetic overweight and obese women. Horm MetabRes 2001;33:602–7.

[36] Kao JT, Wen HC, Chien KL, Hsu HC, Lin SW. A novel genetic variant in the apoli-poprotein A5 gene is associated with hypertriglyceridaemia. Hum Mol Genet2003;12:2533–9.

[37] Talmud PJ, Hawe E, Martin S, et al. Relative contribution of variation within theAPOC3/A4/A5 gene cluster in determining plasma triglycerides. Hum Mol Genet2002;11:3039–46.

[38] Johansen CT, Kathiresan S, Hegele RA. Genetic determinants of plasma triglycerides.J Lipid Res 2011;52:189–206.

[39] Zheng XY, Zhao SP, YanH. The role of apolipoprotein A5 in obesity and themetabolicsyndrome. Biol Rev 2013;88:490–8. http://dx.doi.org/10.1111/brv.12005.

[40] Kim JY, KimOY, Paik JK, Lee SH, Lee JH. Association of apolipoprotein A-V concentrationwith apolipoprotein A5 gene−1131T N C polymorphism and fasting triglyceride levels.J Clin Lipidol 2013;7:94–101.

[41] Sokup A, Goralczyk B, Goralczyk K, Rosc D. Triglycerides as an early pathophysiolog-ical marker of endothelial dysfunction in nondiabetic women with a previous histo-ry of gestational diabetes. Acta Obstet Gynecol Scand 2012;91:182–8. http://dx.doi.org/10.1111/j.1600-0412.2011.01289.x.

[42] Matsumoto S, Gotoh N, Hishinuma S, et al. The role of hypertriglyceridemia in thedevelopment of atherosclerosis and endothelial dysfunction. Nutrients2014;6:1236–50. http://dx.doi.org/10.3390/nu6031236.

[43] Hubacek JA, Pitha J, Adamkova Z, Skodova Z, Lanska V, Poledne R. Apolipoprotein Eand apolipoprotein CI polymorphisms in the Czech population: almost completelinkage disequilibrium of the less frequent alleles of both polymorphisms. PhysiolRes 2003;52:195–200.

[44] Khan TA, Shah T, Prieto D, Zhang W, Price J, Fowkes GR. Apolipoprotein E genotype,cardiovascular biomarkers and risk of stroke: systematic review and meta-analysisof 14,015 stroke cases and pooled analysis of primary biomarker data from up to60,883 individuals. Int J Epidemiol 2013;42(2):475–92. http://dx.doi.org/10.1093/ije/dyt034.

[45] Nagato LC, de Souza Pinhel MA, de Godoy JM, Souza DR. Association of ApoE geneticpolymorphisms with proximal deep venous thrombosis. J Thromb Thrombolysis2012;33:116–9. http://dx.doi.org/10.1007/s11239-011-0639-9.