Clinical Biochemistry 47 (2014) 398–403

Contents lists available at ScienceDirect

Clinical Biochemistry

j ourna l homepage: www.e lsev ie r .com/ locate /c l inb iochem

Association of glutathione-S-transferase gene polymorphism andlipoprotein subclasses in hemodialysis patients

Jelena Vekic a,⁎, Aleksandra Zeljkovic a, Zorana Jelic-Ivanovic a, Tatjana Damjanovic b, Sonja Suvakov c,Marija Matic c, Ana Savic-Radojevic c, Tatjana Simic c, Vesna Spasojevic-Kalimanovska a, Tamara Gojkovic a,Slavica Spasic a, Nada Dimkovic b,d

a Department of Medical Biochemistry, Faculty of Pharmacy, University of Belgrade, Belgrade, Serbiab Clinical Department for Renal Diseases, Zvezdara University Medical Center, Belgrade, Serbiac Institute of Medical and Clinical Biochemistry, Faculty of Medicine, University of Belgrade, Belgrade, Serbiad Faculty of Medicine, University of Belgrade, Belgrade, Serbia

⁎ Corresponding author at: Department of Medical BioUniversity of Belgrade, Vojvode Stepe 450, P.O. Box 146,39 72 840.

E-mail address: jelena.vekic@pharmacy.bg.ac.rs (J. Vek

0009-9120/$ – see front matter © 2013 The Canadian Sochttp://dx.doi.org/10.1016/j.clinbiochem.2013.11.011

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 9 July 2013Received in revised form 14 November 2013Accepted 17 November 2013Available online 27 November 2013

Keywords:Glutathione-S-transferase polymorphismSmallDense LDLHDL subclassesEnd-stage renal diseaseCardiovascular disease risk

Objectives: End-stage renal disease (ESRD) is characterized by profound dyslipidemia and enhancedoxidative stress. The patients also show evidence of exhausted and/or deficient anti-oxidative defense enzymes,one of them being glutathione-S-transferase (GST). This study investigates relationship between GST genepolymorphism and low-density lipoprotein (LDL) and high-density lipoprotein (HDL) subclasses in ESRD.

Design and methods: GSTM1, T1, and P1 genotypes were determined by polymerase chain reaction–restriction fragment length polymorphism in 160 patients undergoing hemodialysis. LDL and HDL subclasseswere separated by gradient gel electrophoresis and biochemical parameters were measured by routine laboratorymethods.

Results: GSTM1-positive patients had higher proportion of small, dense LDL III particles than those with GSTM1-null genotype (P b 0.05). Similarly, GSTP1-Ile/Ile patients had higher proportion of LDL III (P b 0.05), but more HDL2b and less HDL 3a particles than GSTP1-Ile/Val and Val/Val carriers (P b 0.05). LDL subclass distribution in smokerswith GSTM1-null genotype was shifted towards smaller particles, as compared to GSTM1-positive and GSTM1-null

non-smokers. Smokers with GSTP1-Ile/Val and Val/Val genotypes had smaller LDL size than their non-smokingcounterparts (P b 0.05). Both smokers and non-smokers with GSTP1 Ile/Ile genotype had more LDL III particlesthan non-smokers carrying Val allele. Non-smokers with GSTP1 Ile/Ile genotype had more HDL 2b subclasses thannon-smokers with GSTP1-Ile/Val and Val/Val (P b 0.05), but less HDL 3a particles than smokers with GSTP1-Ile/Valand Val/Val genotypes (P b 0.05). GSTT1 gene polymorphism had no effect on lipoprotein subclass distributions.

Conclusions: Our results demonstrate significant associations between low activity GST genotypes andproatherogenic lipoprotein particles in hemodialysis patients which might further increase their cardiovasculardisease risk.

© 2013 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.

Introduction

Progressive renal failure is associated with considerable risk ofcardiovascular disease (CVD) development. However, it is also firmlyestablished that prevention of CVD by evaluating traditional risk factorsis not sufficient for majority of end-stage renal disease (ESRD) patients[1]. Therefore, the interest in improving CVD prevention stimulatedresearch of novel risk factors that could additionally explain such excessof cardiovascular events associated with ESRD. Abnormalities of serumlipid levels, being important CVD risk factors, are typically seen inESRD. Common attributes of dyslipidemia are altered distributions

chemistry, Faculty of Pharmacy,Belgrade, Serbia. Fax: +381 11

ic).

iety of Clinical Chemists. Published b

of low-density (LDL) and high-density (HDL) lipoprotein particles.A growing body of evidence suggests an independent predictive roleof small, dense LDL and small HDL particles in CVD development [2].Our recent data indicate an importance of LDL and HDL subclass assess-ment in patients undergoing hemodialysis [3], and also even after renaltransplantation [4].

Development of CVD has multifactorial etiology and a polygenicbasis. To date, a large scale of candidate genes with potential associationwith CVD has been identified. Glutathione-S-transferases are a super-family of polymorphic enzymes that play an important role in detoxifi-cation by catalyzing the conjugation of hydrophobic compounds withreduced glutathione [5]. Human GSTs are categorized into four mainclasses: alpha (GSTA), mu (GSTM1), theta (GSTT1) and pi (GSTP1). Ithas been firmly established that genetic polymorphism of GST affectsenzyme activity. As a result, deletions in GSTT1 and GSTM1 genes (nullgenotypes) are accompanied by lack of enzymatic activity [6]. In the

y Elsevier Inc. All rights reserved.

399J. Vekic et al. / Clinical Biochemistry 47 (2014) 398–403

case ofGSTP1, the nucleotide transition of A → G leads to Ile → Val sub-stitution and such a new allele has altered activity towards substratescompared to wild-type allele [6]. Our previous data suggest that GSTpolymorphism is associated with increased risk of kidney and urinarybladder tumors [7], as well as of ESRD development [8]. Since theenzyme is also involved in the biotransformation of several compoundsrecognized as risk factors for CVD, association between geneticpolymorphism of GST and risk of CVD has been recently proposed. Itseems that individuals with GSTM1 and/or GSTT1-null genotypes areat a higher risk of CVD [9,10]. However, other studies have yieldedconflicting results [11]. In addition to genetic predisposition, numerousenvironmental factors may further contribute to CVD susceptibility.In accordance, the observed link between GST polymorphism and CVDwas further strengthened in smokers lacking GSTM1 or GSTT1 genes[12,13]. However, there is no data on relationship between GST genepolymorphisms and smoking with respect to potential CVD risk inESRD patients.

Previous studies of genetic polymorphism affecting LDL and HDLsubclass distributions have focused almost exclusively on the enzymesthat participate in lipoprotein metabolism. However, besides dyslipid-emia, oxidative stress is deeply involved in the pathophysiology ofCVD [14]. Our previous data indicate that oxidative stress is associatedwith impaired lipoprotein subclass distributions, even in the absenceof clinical symptoms of atherosclerosis [15]. Hemodialysis patients areparticularly prone to oxidative stress, as a consequence of the diseaseitself and the dialysis process [16]. Our recent results showed thatsusceptibility to oxidative stress in hemodialysis patients is stronglyinfluenced by GST genotype. Specifically, patients with GSTA1, GSTM1,GSTP1 and GSTT1-null/low-activity genotypes, exhibit significantlyhigher level of oxidative damage [8]. Since GST enzymes play an impor-tant role in the anti-oxidative defense system, we hypothesized thatGST gene polymorphismmay also affect LDL and HDL particle distribu-tions in this category of patients.

The aim of this study was to evaluate the effects of GSTM1, T1 and P1gene polymorphisms, aswell as their possible interactionwith smoking,on LDL and HDL subclass distributions in patients undergoing regularhemodialysis.

Materials and methods

Patients

The original sample size included 199 patients undergoing hemodial-ysis treatment for 12–15 h per week in two dialysis facilities in Belgrade(Center for Renal Diseases, University Medical Center Zvezdara andDepartment of Nephrology and Haemodialysis, University TeachingHospital Zemun). A detailed description of the study protocol has beenpublished elsewhere [8]. Patients were treated with single-use dialyzersequippedwith low- and high-flux polysulfonemembranes. The efficien-cy of dialysis was estimated by calculation of Kt/V (K, clearance of urea ofthe dialyzer, t, time of dialysis and V, volume of distribution of urea). Thecauses of ESRDwere hypertensive nephrosclerosis (n = 74), glomerulo-nephritis (n = 20), diabetic nephropathy (n = 24), polycystic renaldisease (n = 16), pyelonephritis (n = 17), Balkan endemic nephro-pathy (n = 7) and obstructive nephropathy (n = 2). All patients werestable, with hemodialysis vintage N3 months before the study. At base-line, essential information concerning the patients was collected. Thesedata included age, gender, hemodialysis duration and smoking habits.The study protocol also included measurement of systolic and diastolicpressure, as well as height and weight for the determination of bodymass index (BMI). Patients with malignancy or infectious diseaseswere not included in the study in order to avoid potential interferencewith lipid status parameters. Current cigarette users were classified assmokers, while the non-smoker category comprised non-users and ex-users, if they quit smoking at least 6 months prior to the study. Accordingto the study protocol [8], patients did not receive any antioxidant

therapy, to avoid effects on oxidative stress status. We excluded 30(15.1%) patients on statin therapy [8], since such lipid-lowering treat-ment affects lipoprotein subclass distribution. In addition, nine patientshad insufficient sample volume for lipoprotein subclass determinationand were not included in the study. Accordingly, the final studygroup comprised 160 patients (91 men and 69 women, mean age60.5 ± 12.4 years). The study was planned and executed followingthe ethical guidelines of the Helsinki Declaration. According to insti-tutional guidelines, the study protocol was approved by the local in-stitutional ethics review board and informed consent was obtainedfrom each subject involved in the study.

Laboratory analyses

Blood samples were collected into evacuated tubes containing EDTAand serum sample tubes after a 12-h fasting period. Plasma and serumwere separated by immediate centrifugation at 1500 · g for 10 minat 4 °C. Aliquots of each sample were stored at −80 °C. The sampleswere thawed immediately before analyses.

Serum urea, creatinine and total protein concentrations weremeasured by standard laboratory methods. Total cholesterol (TC) andtriglyceride (TG) concentrationswere assayed using enzymaticmethodsusing an ILab 600 analyzer (Instrumentation Laboratory, Lexington,MA, USA). The concentration of HDL-C was measured using the sameenzymatic method after precipitation of the plasma with phosphotung-stic acid in the presence of magnesium ions. The concentration ofLDL-C was determined using a direct homogeneous assay (OlympusDiagnostica GmbH, Hamburg, Germany).

Plasma LDL andHDL particleswere separated using an electrophore-sismethod (Amersham Pharmacia Biotech, Vienna, Austria)with ImageQuant software (version 5.2; 1999;Molecular Dynamics, Sunnyvale, CA,USA) previously described by Rainwater et al. [17]. A detailed descrip-tion of the procedure has been published elsewhere [18]. The sizes ofLDL ≤ 25.5 nm andHDL ≤ 8.8 nmwere the criteria used for the defini-tion of smaller, denser LDL and HDL particles (i.e., LDL B and HDL 3 phe-notypes) [18]. The relative content of each LDL and HDL subclass wasestimated by determining the areas under the peaks of densitometryscans [19,20].

Genomic DNA was isolated from whole blood using the QIAGENQIAamp kit (Qiagen, Inc., Chatsworth, CA). GSTM1 genotyping wasperformed by multiplex PCR [21]. Used primers were GSTM1 forward:5′-GAACTCCCTGAAAAGCTAAAGC-3′ and GSTM1 reverse: 5′ GTTGGGCTCAAATATACGGTGG-3′. Exon 7 of CYP1A1 gene was co-amplifiedand used as an internal control using the following primers: CYP1A1forward: 5′-GAACTGCCACTTCAGCTGTCT-3′ and CYP1A1 reverse:5′-CAGCTGCATTTGGAAGTGCTC-3′. The presence of GSTM1 activegenotype was detected by the band at 215 bp, since the assaydoes not distinguish heterozygous or homozygous wild-type geno-types. GSTT1 genotyping was performed by multiplex PCR [21].Used primers were GSTT1-forward: 5′-TTCCTTACTGGTCCTCACATCTC-3′ and GSTT1-reverse: 5′-TCACGGGATCATGGCCAGCA-3′. Theassay does not distinguish between heterozygous or homozygouswild-type genotypes; therefore the presence of 480 bp bands wasindicative for GSTT1 active genotype. GSTP1 Ile105Val polymor-phism was analyzed using PCR-RFLP method by Harries et al. [22].Used primers were: GSTP1 Ile105Val forward: 5′-ACCCCAGGGCTCTATGGGAA-3′ and GSTP1 Ile105Val reverse: 5′-TGAGGGCACAAGAAGCCCCT-3′. Presence of restriction site resulting in two fragments(91 bp and 85 bp) indicated mutant allele (Val/Val) while if Ile/Valpolymorphism incurred, it resulted in one more fragment of 176 bp.

Statistical analyses

Data are shown as mean ± standard deviation (SD) for normallydistributed continuous variables and as relative or absolute frequenciesfor categorical variables. Since the distribution of TG was skewed, data

400 J. Vekic et al. / Clinical Biochemistry 47 (2014) 398–403

were logarithmically transformed before analysis and presentedas median and interquartile range (IQR). Differences in continuousvariables were analyzed using the Student's t-test and ANOVA withthe LSD post hoc test for subgroup differences. Analysis of categoricalvariables was performed using Chi-square tests for contingency tables.

Results

Table 1 presents clinical and laboratory characteristics of thepatients. Homozygote deletion of GSTM1 gene was observed in 95(59.4%) patients, while GSTT1-null genotype had 50 (31.2%) studyparticipants. The Ile/Ile genotype of GSTP1 was identified in 58 (36.2%)patients, 67 (41.9%) had Ile/Val genotype, whereas 35 (21.9%) patientswere Val/Val homozygotes. The prevalence of LDL B and HDL 3 pheno-types was 28.1% and 30.6%, respectively.

Lipid status parameters and distributions of LDL and HDL subclassesaccording to GST gene polymorphisms are shown in Table 2. Geneticpolymorphism of GST had no significant effects on serum lipid parame-ters in examined group.We found that GSTM1-positive subjects had sig-nificantly higher relative proportion of small, dense LDL III subclassesthan individuals with GSTM1-null genotype. The proportion of LDL IIIsubclasses was also increased in the carriers of GSTP1-Ile/Ile genotypecomparing to GSTP1 Ile/Val and Val/Val patients. On the other hand,GSTP1-Ile/Ile homozygotes hadmore HDL 2b and less HDL 3a subclasseswhen compared with the GSTP1-Ile/Val and Val/Val genotype patients.Although GSTT1 genotype had no effect on distributions of LDL sub-classes, we found an increased prevalence of HDL 3 phenotype amongGSTT1-null carriers.

Next we analyzed the effect of interaction between GST gene poly-morphism and smoking status on LDL and HDL subclass distributions.Our results have shown that smokers with GSTM1-null genotype hadhigher proportion of the smallest LDL IV subclasses and lower percentageof LDL II particles than GSTM1-positive smokers, and also than non-smokers regardless of GSTM1 genotype. We found no differences inHDL subclass distributions among the examined groups (Table 3).

The results of interaction between GSTP1 genotype and smokingstatus on LDL and HDL subclass distributions are presented in Table 4.We found that smokers with GSTP1-Ile/Val and Val/Val genotypes hadsmaller mean LDL size than their non-smoking counterparts. Bothsmokers and non-smokers with GSTP1 Ile/Ile genotype had more LDLIII subclasses than non-smokers with GSTP1-Ile/Val and Val/Val geno-types. Non-smoking GSTP1-Ile/Ile homozygotes had higher proportionof HDL 2b subclasses than non-smokers with GSTP1-Ile/Val and Val/Valand lower proportion of HDL 3a subclasses than smokers withGSTP1-Ile/Val and Val/Val genotypes. There was no effect of inter-action between GSTT1 genotype and smoking on lipoprotein sub-class distributions (data not shown).

Table 1Clinical and laboratory characteristics of the patients.

Parameter Mean ± SD

Age (years) 60.5 ± 12.4Gender, male (%) 56.9BMI (kg/m2) 24.4 ± 4.2Systolic blood pressure (mm Hg) 142.4 ± 18.6Diastolic blood pressure (mm Hg) 82.4 ± 8.7Smokers (%) 25.9Hemodialysis duration (years) 6.1 ± 4.3Kt/V 1.3 ± 0.3Creatinine (μmol/L) 860.9 ± 249.6Urea (mmol/L) 24.2 ± 5.3Total protein (g/L) 66.6 ± 5.9

BMI, body mass index.

Discussion

To date, the atherogenic role of small, dense LDL and small HDL par-ticles has been documented in a large scale of clinical and experimentalstudies. It has been demonstrated that smaller LDL particles easily pen-etrate to arterial intima, reside longer in subendothelium and are moreprone to oxidation, as compared to their larger counterparts [2]. Also,smaller HDL particles, even if they are considered as essentially protec-tive, could have decreased anti-atherogenic capacity in dyslipidemia,inflammation and enhanced oxidative stress [23], all regularly seen inESRD [16]. Abnormalities in lipid homeostasis are commonly seen inchronic kidney disease patients [24]. Moreover, unfavorable lipoproteinprofile is present early in the course of impaired kidney function. Inaccordance, Reiman and colleagues [25] investigated LDL subclass dis-tribution in chronic kidney disease patients of different etiologies, in-cluding a group of predialysis patients suffering glomerulonephritis,polycystic kidney disease and obstructive nephropathy. The authors[25] found that predialysis group of patients had increased small,dense LDL particles, suggesting that the disease itself contributes toadverse lipoprotein subclass distribution. However, to the best of ourknowledge, no previous study investigated the impact of particularrenal disease etiology on lipoprotein subclass profile and future studiesare needed to resolve this specific issue.

GST gene polymorphism affects enzymatic activity and is believed tomodulate CVD risk [6]. The current study is the first to demonstrate sig-nificant associations between GST gene polymorphism and lipoproteinsubclasses in patients on hemodialysis. According to recent meta-analysis of Wang et al. [26], GSTM1-null polymorphism contributes tohigher CVD risk, whereas the GSTT1 polymorphism seems unrelated.Recently Lin et al. [27] reported that GSTM1-positive patients on hemo-dialysis had significantly better overall survival than thosewith GSTM1-null genotype. However, we found that GSTM1-positive patients had sig-nificantly higher proportion of small, dense LDL III subclasses than thosewith GSTM1-null genotype (Table 2). If we appreciate that small, denseLDL particles are more prone to oxidation and, therefore, have greateratherogenic potential than their larger counterparts [2], we couldspeculate that those patients could have enhanced CVD risk. A possibleexplanation could be the fact that the level of glutathione continuallydecreases as chronic kidney disease progresses [28]. Hence, even ifGST participates in removal of cytotoxic aldehydes produced duringlipid peroxidation [6], there is a possibility for adverse modification ofaccumulated small, dense LDL particles in GSTM1-positive subjects inspite of conserved enzymatic activity. In support to the previous state-ment, our analysis of GSTP1 polymorphism revealed similar results,since we found that carriers of GSTP1 Ile/Ile genotype had significantlyhigher proportion of small, dense LDL III subclasses than GSTP1 Ile/Valand Val/Val patients (Table 2). Yet, the same analysis revealed an oppo-site association of GSTP1 genotypes and HDL subclasses. Namely, ourGSTP1-Ile/Ile homozygotes hadmore HDL 2b and less HDL 3a subclassesthan the carriers of GSTP1-Ile/Val and Val/Val genotypes (Table 2). Thisfavorable HDL distribution in the carriers of GSTP1-Ile/Ile genotype sug-gests an importance of HDL's antiatherogenic potential in the conditionsassociated with exacerbated oxidative stress and exhausted plasmaglutathione levels.

To date, epidemiological studies have indicated that GST polymor-phisms are associated with increased risk for cancers among smokers,but there are still some controversies about the relationship betweenGST polymorphism and CVD risk in smokers. According to resultsof Kim et al. [29], smokers with GSTM1-null genotype have twice asmuch chance for coronary atherosclerosis than GSTM1-positive non-smokers. In the present study, we found that smokers with GSTM1-null genotypehad increasedproportion of the smallest LDL IV subclassesand reduced percentage of LDL II particles when compared to GSTM1-positive and GSTM-null non-smokers (Table 3). These results clearlyconfirm adverse effects of smoking on LDL metabolism [30]. Moreover,they also emphasize the potential for further aggravation of CVD

Table 2Lipid status parameters and LDL and HDL subclasses according to GST genotypes.

GSTM1 GSTT1 GSTP1

Positive (n = 65) Null (n = 95) Positive (n = 110) Null (n = 50) Ile/Ile (n = 58) Ile/Val + Val/Val (n = 102)

TC (mmol/L) 4.8 ± 1.1 4.6 ± 1.1 4.8 ± 1.2 4.5 ± 0.9 4.6 ± 1.2 4.7 ± 1.1LDL-C (mmol/L) 2.7 ± 0.8 2.5 ± 0.7 2.7 ± 0.8 2.5 ± 0.6 2.6 ± 0.9 2.6 ± 0.7HDL-C (mmol/L) 1.2 ± 0.4 1.8 ± 0.3 1.2 ± 0.4 1.2 ± 0.3 1.1 ± 0.3 1.2 ± 0.4TG (mmol/L)a 1.8 (1.3–2.7) 1.8 (1.1–2.6) 1.8 (1.3–2.5) 1.8 (1.2–2.8) 1.8 (1.2–2.8) 1.8 (1.3–2.5)LDL size (nm) 26.2 ± 1.4 26.6 ± 1.5 26.4 ± 1.5 26.6 ± 1.3 26.3 ± 1.4 26.5 ± 1.5

LDL subclasses (%)LDL I 21.1 ± 6.0 22.3 ± 6.9 21.8 ± 7.4 21.8 ± 4.5 21.1 ± 5.7 25.3 ± 7.1LDL II 29.3 ± 7.0 27.9 ± 6.9 28.3 ± 6.8 28.8 ± 7.2 28.8 ± 6.2 28.3 ± 7.3LDL III 23.1 ± 6.0 21.4 ± 4.9b 22.3 ± 5.8 21.6 ± 4.6 23.4 ± 5.3 21.3 ± 5.4c

LDL IV 26.5 ± 7.4 28.4 ± 8.3 27.5 ± 8.1 27.8 ± 7.7 26.7 ± 7.2 28.1 ± 8.4LDL B phenotype (%) 32.8 25 30 24 34.5 24.5HDL size (nm) 9.5 ± 1.1 9.5 ± 0.9 9.49 ± 1.02 9.5 ± 1.1 9.7 ± 1.1 9.4 ± 0.9

HDL subclasses (%)HDL 2b 41.9 ± 9.9 42.9 ± 10.8 42.1 ± 9.9 43.3 ± 11.6 44.7 ± 9.9 41.2 ± 10.5c

HDL 2a 22.2 ± 3.9 21.7 ± 4.4 22.0 ± 4.3 21.7 ± 4.1 21.5 ± 3.8 22.1 ± 4.5HDL 3a 15.8 ± 3.7 15.5 ± 4.1 15.8 ± 3.7 15.3 ± 4.6 14.9 ± 3.2 16.0 ± 4.3c

HDL 3b 9.4 ± 3.1 9.5 ± 3.6 9.6 ± 3.1 9.3 ± 3.9 8.9 ± 3.2 9.8 ± 3.4HDL 3c 10.7 ± 4.6 10.4 ± 4.8 10.6 ± 4.5 10.4 ± 5.2 9.9 ± 5.0 10.9 ± 4.5

HDL 3 phenotype (%) 29.7 31.2 25.2 40.4d 29.5 31.3

Continuous variables are compared by Student's t-test and categorical variables with Chi-square test. GST, glutathione-S-transferase; BMI, body mass index; TC, total cholesterol; LDL-C,low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; TG, triglycerides.

a Data are presented as median (IQR).b Significantly different from GSTM1-positive group (P b 0.05).c Significantly different from GSTP1 Ile/Ile group (P b 0.05).d Significantly different from GSTT1-positive group (P b 0.05).

401J. Vekic et al. / Clinical Biochemistry 47 (2014) 398–403

risk in this category of patients. Namely, it has been demonstratedthat smoking enhances free radical production [31]. As previouslymentioned, once the patient undergoes hemodialysis, the level ofglutathione dramatically decreases [28]. In this scenario, the lack ofGST enzyme activity in GSTM1-null smoking patients will contribute toaugmented oxidative stress and this might, in turn, increase LDL oxida-tion. Finally, our finding that among smokers, GSTM1-positive patientshave more favorable LDL subclass distribution than those with GSTM-null genotype, suggests potential etiological role of GSTM1 polymor-phism for the predominance of small, dense LDL particles. Similarly,non-smokers with GSTP1 Ile/Ile genotype had more HDL 2b subclassesthan non-smokers with GSTP1-Ile/Val and Val/Val and less HDL 3aparticles than smokers GSTP1-Ile/Val and Val/Val genotypes (Table 4).Based on the observed results, we assume that, apart from the established

Table 3Interaction of GSTM1 genotype and smoking status on LDL and HDL subclasses.

GSTM1-positive/non-smoker (n = 50) GSTM1-null/non-smoker (

LDL size (nm) 26.3 ± 1.3 26.7 ± 1.4

LDL subclasses (%)LDL I 21.1 ± 6.4 22.7 ± 7.5LDL II 28.9 ± 7.1b 28.9 ± 7.5b

LDL III 22.9 ± 6.2 21.1 ± 4.7LDL IV 27.1 ± 7.6b 27.2 ± 9.0b

LDL B phenotype (%) 32 19.4HDL size (nm) 9.5 ± 1.1 9.4 ± 0.9

HDL subclasses (%)HDL 2b 42.8 ± 9.6 42.6 ± 10.8HDL 2a 21.6 ± 3.8 21.9 ± 4.7HDL 3a 15.5 ± 3.6 15.3 ± 3.9HDL 3b 9.4 ± 3.2 9.6 ± 3.6HDL 3c 10.8 ± 4.5 10.6 ± 5.1

HDL 3 phenotype (%) 28.6 34.3

GST, glutathione-S-transferase; LDL, low-density lipoprotein; HDL, high-density lipoprotein.a Continuous variables are compared by ANOVA with LSD post hoc test and categorical variab Significantly different from GSTM1-null/smoker group (P b 0.05).

role of smoking [32], GSTP1 polymorphism could also interact in HDLremodeling.

Previous studies revealed that GSTT1-null genotype influenceshigher risk for ESRD development [33,34]. We have recently showedthat ESRD patients with combined GSTM1/GSTT1-null genotypes exhibitsignificantly higher level of oxidative damage [8]. Data from currentstudy do not support an association of GSTT1 genotypes with LDL parti-cle distributions, probably due to a low prevalence of GSTT1-null geno-type in the studied group. Still, it is important to note that patientswith HDL 3 phenotype were more prevalent among GSTT1-null carriers(Table 2). Our previous results indicate that hemodialysis patients withHDL 3 phenotype have reduced survival than those with large HDLs [3].As GSTT1-null patients have compromised enzymatic activity [6], weassume that both conditions are able to exacerbate patient's CVD risk,

n = 68) GSTM1-positive/smoker (n = 15) GSTM1-null/smoker (n = 27) Pa

25.9 ± 1.5 26.2 ± 1.7 0.226

21.2 ± 4.6 20.6 ± 5.2 0.41730.5 ± 6.6b 25.6 ± 4.3 0.08323.9 ± 5.6 22.6 ± 5.1 0.15124.4 ± 6.7b 31.2 ± 5.5 b0.0535.7 40.7 0.1479.5 ± 1.1 9.6 ± 1.0 0.884

38.9 ± 10.5 43.3 ± 11.1 0.58224.4 ± 3.6 21.3 ± 3.9 0.10416.8 ± 4.1 16.1 ± 4.9 0.5609.3 ± 2.7 9.4 ± 2.8 0.97410.6 ± 4.9 10.0 ± 4.1 0.92135.7 25.9 0.814

bles with Chi-square test.

Table 4Interaction of GSTP1 genotype and smoking status on LDL and HDL subclasses.

GSTP1-Ile/Ile, non-smoker(n = 41)

GSTP1-Ile/Val + Val/Val, non-smoker(n = 77)

GSTP1-Ile/Ile, smoker(n = 17)

GSTP1-Ile/Val + Val/Val, smoker(n = 25)

Pa

LDL size (nm) 26.2 ± 1.4 26.7 ± 1.3b 26.4 ± 1.4 25.9 ± 1.7 0.090

LDL subclasses (%)LDL I 20.7 ± 6.1 22.7 ± 7.5 21.2 ± 4.3 20.6 ± 5.4 0.317LDL II 29.2 ± 6.4 28.8 ± 7.8 28.0 ± 6.1 26.8 ± 5.5 0.561LDL III 23.4 ± 5.9c 21.0 ± 5.1 23.8 ± 3.8c 22.5 ± 6.1 0.063LDL IV 26.7 ± 7.7 27.4 ± 8.8 26.9 ± 6.5 30.1 ± 6.7 0.378

LDL B phenotype (%) 35 19.5 35.3 41.7 0.096HDL size (nm) 9.5 ± 1.1 9.4 ± 0.9 9.8 ± 0.9 9.4 ± 1.1 0.402

HDL subclasses (%)HDL 2b 45.5 ± 8.7c 41.2 ± 10.7 42.1 ± 12.1 41.4 ± 10.2 0.186HDL 2a 21.1 ± 3.5 22.2 ± 4.7 22.9 ± 3.9 22.1 ± 4.1 0.423HDL 3a 14.6 ± 2.9b 15.8 ± 4.0 15.7 ± 3.7 16.8 ± 5.1 0.175HDL 3b 8.9 ± 3.3 9.8 ± 3.7 9.0 ± 3.1 9.6 ± 2.5 0.574HDL 3c 9.9 ± 4.9 11.1 ± 4.8 10.3 ± 5.4 10.1 ± 3.5 0.570

HDL 3 phenotype (%) 30 32.9 31.6 27.3 0.962

GST, glutathione-S-transferase; LDL, low-density lipoprotein; HDL, high-density lipoprotein.a Continuous variables are compared by ANOVA with LSD post hoc test and categorical variables with Chi-square test.b Significantly different from GSTP1-Ile/Val + Val/Val, smoker group (P b 0.05).c Significantly different from GSTP1-Ile/Val + Val/Val, non-smoker group (P b 0.05).

402 J. Vekic et al. / Clinical Biochemistry 47 (2014) 398–403

but future investigations, with a much larger study group, are neededto confirm this observation.

According to the study protocol [8] none of our patients receivedanti-oxidative treatment, due to previously reported effects of suchsupplementation on oxidative stress status in hemodialysis patients.Namely, a recent systematic review by Coombes and Fassett [35]revealed that majority of examined studies showed a decrease inbiomarkers of oxidative stress following antioxidant therapy. SinceGST enzymes play an important role in the anti-oxidative defense sys-tem, itwas obligatory to avoid individuals on anti-oxidative supplemen-tation. The results of the studies investigating the influence of anti-oxidative treatment on lipid profile are ambiguous. Nevertheless,Abdollahzad et al. [36] demonstrated that anti-oxidative supplementa-tion decreased lipid peroxidation and improved lipid profile in hemodi-alysis patients. Regarding lipid-lowering medication, it has been welldocumented that such therapy, in general, affects lipid profile andlipoprotein subclass distribution. Furthermore, it has been demonstratedthat concomitant use of anti-oxidative supplements with lipid-loweringmedication may produce adverse effects. In a study of Brown et al. [37]when antioxidants were combinedwith lipid-lowering treatment, bene-fits on serum lipid profile tended to diminish as compared with thoseachieved with lipid-lowering agents alone. Finally, according to a recentmeta-analysis by the Lipid and Blood Pressure Meta-Analysis Collabora-tion Group [38], the available studies have reported the benefit of statinuse in chronic kidney disease patients. However, in patients on dialysisstatins showed less beneficial effect and tend to be less effective withlonger duration of therapy [38].

Our study has some limitations. At first, presented findings arederived from the patients of Serbian origin, which could limit generali-zation of our results to other ethnic groups. Also, despite a fairly largenumber of subjects included in the study, in certain subgroups thenumber of patients was rather small for reaching statistical significance.Finally, the cross-sectional nature of the study did not allow us toinvestigate a causal relationship between GST gene polymorphism andalterations of lipoprotein profile in ESRD with development of CVD inlater life. Nevertheless, this study may offer some basic observationsthat could form the substrate for future research.

Conclusions

In conclusion, the current study is the first to describe associations ofGST gene polymorphisms with lipoprotein subclasses in patients onhemodialysis. In addition, the results presented herein demonstrate

significant gene-smoking interactions on adverse LDL and HDL subclassdistributions in ESRD. Future studies will shed light on the observedassociations and their possible consequences on CVD risk in this categoryof patients.

Conflict of interest

The authors have no conflict of interest to declare.

Acknowledgment

This work was supported by a grant from theMinistry of Education,Science and Technological Development, Republic of Serbia (Project No.175035) and by the European Cooperation in Science and Technology(COST) BM0904 Action.

References

[1] Weiner DE, Tighiouart H, Elsayed EF, Griffith JL, Salem DN DN, Levey AS, et al.The Framingham predictive instrument in chronic kidney disease. J Am Coll Cardiol2007;50:217–24.

[2] Rizzo M, Berneis K, Zeljkovic A, Vekic J. Should we routinely measure low-densityand high-density lipoprotein subclasses. Clin Lab 2009;55:421–9.

[3] Vekic J, Zeljkovic A, Bogavac-Stanojevic N, Jelic-Ivanovic Z, Spasojevic-KalimanovskaV, Simic-Ogrizovic S, et al. Cox proportional hazard model analysis of survival inend-stage renal disease patients with small-sized high-density lipoprotein particles.Clin Biochem 2011;44:635–41.

[4] Zeljkovic A, Vekic J, Spasojevic-Kalimanovska V, Jelic-Ivanovic Z, Peco-Antic A, KosticM, et al. Characteristics of low-density and high-density lipoprotein subclasses inpediatric renal transplant recipients. Transpl Int 2011;24:1094–102.

[5] Wilce M, Parker M. Structure and function of glutathione S-transferase. BiochimBiophys Acta 1994;1205:1–18.

[6] Strange RC, Jones PW, Fryer AA. Glutathione S-transferase: genetics and role intoxicology. Toxicol Lett 2000;112–113:357–63.

[7] Simic T, Savic-Radojevic A, Pljesa-Ercegovac M, Matic M, Mimic-Oka J. GlutathioneS-transferases in kidney and urinary bladder tumors. Nat Rev Urol 2009;6:281–9.

[8] Suvakov S, Damjanovic T, Stefanovic A, Pekmezovic T, Savic-Radojevic A,Pljesa-Ercegovac M, et al. Glutathione S-transferase A1, M1, P1 and T1 null or low-activity genotypes are associated with enhanced oxidative damage amonghaemodialysis patients. Nephrol Dial Transplant 2013;28:202–12.

[9] Nomani H, Mozafari H, Ghobadloo SM, Rahimi Z, Raygani AV, Rahimi MA, et al.The association between GSTT1, M1, and P1 polymorphisms with coronary arterydisease in Western Iran. Mol Cell Biochem 2011;354:181–7.

[10] Ramprasath T, Senthil Murugan P, Prabakaran AD, Gomathi P, Rathinavel A, SelvamGS. Potential risk modifications of GSTT1, GSTM1 and GSTP1 (glutathione-S-transferases) variants and their association to CAD in patients with type-2 diabetes.Biochem Biophys Res Commun 2011;407:49–53.

[11] Wilson M, Grant P, Hardie L, Wild C. Glutathione-S-transferase M1-null genotype isassociated with a decreased risk of myocardial infarction. FASEB J 2000;14:791–6.

403J. Vekic et al. / Clinical Biochemistry 47 (2014) 398–403

[12] Wang LS, Tang JJ, Tang NP, Wang MW, Yan JJ, Wang QM, et al. Association of GSTM1and GSTT1 gene polymorphisms with coronary artery disease in relation to tobaccosmoking. Clin Chem Lab Med 2008;46:1720–5.

[13] Masetti S, Botto N, Manfredi S, Colombo MG, Riza A, Vassalle C. Interactive effectof the glutathione-S-transferase genes and cigarette smoking on occurrence andseverity of coronary artery risk. J Mol Med 2003;81:488–94.

[14] Rizzo M, Kotur-Stevuljevic J, Berneis K, Spinas G, Rini GB, Jelic-Ivanovic Z, et al.Atherogenic dyslipidemia and oxidative stress: a new look. Transl Res2009;153:217–23.

[15] Vekic J, Kotur-Stevuljevic J, Jelic-Ivanovic Z, Spasic S, Spasojevic-Kalimanovska V,Topic A, et al. Association of oxidative stress and PON1 with LDL and HDL particlesize in middle-aged subjects. Eur J Clin Invest 2007;37:715–23.

[16] Libetta C, Sepe V, Esposito P, Galli F, Dal Canton A. Oxidative stress and inflammation:implications in uremia and hemodialysis. Clin Biochem 2011;44:1189–98.

[17] Rainwater DL, Moore PH, Gamboa IO. Improved method for making nondenaturingcomposite gradient gels for the electrophoretic separation of lipoproteins. J Lipid Res2004;45:773–5.

[18] Vekic J, Topic A, Zeljkovic A, Jelic-Ivanovic Z, Spasojevic-Kalimanovska V. LDL andHDL subclasses and their relationship with Framingham risk score in middle-agedSerbian population. Clin Biochem 2007;40:310–6.

[19] Zeljkovic A, Vekic J, Spasojevic-Kalimanovska V, Jelic-Ivanovic Z, Bogavac-StanojevicN, Gulan B, et al. LDL and HDL subclasses in acute ischemic stroke: prediction ofrisk and short-term mortality. Atherosclerosis 2010;210:548–54.

[20] Zeljkovic A, Spasojevic-Kalimanovska V, Vekic J, Jelic-Ivanovic Z, Topic A,Bogavac-Stanojevic M, et al. Does simultaneous determination of LDL and HDLparticle size improve prediction of coronary artery disease risk? Clin Exp Med2008;8:109–16.

[21] Abdel-Rahman SZ, el-Zein RA, Anwar WA, Au WW. A multiplex PCR procedure forpolymorphic analysis of GSTM1 and GSTT1 genes in population studies. CancerLett 1996;107:229–33.

[22] Harries LW, Stubbins MJ, Forman D, Howard GC, Wolf CR. Identification of geneticpolymorphisms at the glutathione S-transferase Pi locus and associationwith susceptibility to bladder, testicular and prostate cancer. Carcinogenesis1997;18:641–4.

[23] Tsompanidi EM, Brinkmeier MS, Fotiadou EH, Giakoumi SM, Kypreos KE. HDLbiogenesis and functions: role of HDL quality and quantity in atherosclerosis.Atherosclerosis 2010;208:3–9.

[24] Tsimihodimos V, Mitrogianni Z, Elisaf M. Dyslipidemia associated with chronickidney disease. Open Cardiovasc Med J 2011;5:41–8.

[25] Reiman I, Harper L, McPake D, Kendall MJ, Wheeler DC. Low-density lipo-protein subfraction profiles in chronic renal failure. Nephrol Dial Transplant1998;13:2281–7.

[26] Wang J, Zou L, Huang S, Lu F, Lang X, Han L, et al. Genetic polymorphisms ofglutathione S-transferase genes GSTM1, GSTT1 and risk of coronary heart disease.Mutagenesis 2010;25:365–9.

[27] Lin YS, Hung SC, Wei YH, Tarng DC. GST M1 polymorphism associates with DNAõoxidative damage and mortality among hemodialysis patients. J Am Soc Nephrol2009;20:405–15.

[28] Ceballos-Picot I, Witko-Sarsat V, Merad-Boudia M, Nguyen AT, Thévenin M, JaudonMC, et al. Glutathione antioxidant system as a marker of oxidative stress in chronicrenal failure. Free Radic Biol Med 1996;21:845–53.

[29] Kim SJ, Kim MG, Kim KS, Song JS, Yim SV, Chung JH. Impact of glutathione S-transferase M1 and T1 gene polymorphisms on the smoking-related coronaryartery disease. J Korean Med Sci 2008;23:365–72.

[30] Chelland Campbell S, Moffatt RJ, Stamford BA. Smoking and smoking cessation— therelationship between cardiovascular disease and lipoprotein metabolism: a review.Atherosclerosis 2008;201:225–35.

[31] Yarnell JW. Smoking and cardiovascular disease. QJM 1996;89:493–8.[32] Gepner AD, Piper ME, Johnson HM, Fiore MC, Baker TB, Stein JH. Effects of smoking

and smoking cessation on lipids and lipoproteins: outcomes from a randomizedclinical trial. Am Heart J 2011;161:145–51.

[33] Datta SK, Kumar V, Pathak R, Tripathi AK, Ahmed RS, Kalra OP, et al. Associationof glutathione S transferase M1 and T1 gene polymorphism with oxidative stressin diabetic and nondiabetic chronic kidney disease. Ren Fail 2010;32:1189–95.

[34] Yang Y, KaoMT, Chang CC, Chung SY, Chen CM, Tsai JJ, et al. Glutathione S-transferaseT1 deletion is a risk factor for developing end-stage renal disease in diabetic patients.Int J Mol Med 2004;14:855–9.

[35] Coombes JS, Fassett RG. Antioxidant therapy in hemodialysis patients: a systematicreview. Kidney Int 2012;81:233–46.

[36] Abdollahzad H, Eghtesadi S, Nourmohammadi I, Khadem-Ansari M, Nejad-Gashti H,Esmaillzadeh A. Effect of vitamin C supplementation on oxidative stress and lipidprofiles in hemodialysis patients. Int J Vitam Nutr Res 2009;79:281–7.

[37] Brown BG, Zhao XQ, Chait A, Fisher LD, Cheung MC, Morse JS, et al. Simvastatinand niacin, antioxidant vitamins, or the combination for the prevention of coronarydisease. N Engl J Med 2001;345:1583–92.

[38] Nikolic D, Nikfar S, Salari P, Rizzo M, Ray KK, Pencina MJ, et al. Effects of statinson lipid profile in chronic kidney disease patients: a meta-analysis of randomizedcontrolled trials. Curr Med Res Opin 2013;29:435–51.