Genetic variability of glutathione S-transferase enzymes in human populations: Functional inter-ethnic differences in detoxification systems

Gene 512 (2013) 102–107

Contents lists available at SciVerse ScienceDirect

Gene

j ourna l homepage: www.e lsev ie r .com/ locate /gene

Short Communication

Genetic variability of glutathione S-transferase enzymes in human populations:Functional inter-ethnic differences in detoxification systems

Renato Polimanti a, Cinzia Carboni a, Ilenia Baesso a, Sara Piacentini a, Andrea Iorio b,Gian Franco De Stefano a, Maria Fuciarelli a,⁎a Department of Biology, University of Rome “Tor Vergata”, Rome, Italyb Clinical Pathophysiology Center, AFaR, “San Giovanni Calibita” Fatebenefratelli Hospital, Rome, Italy

Abbreviations: GSTs, Glutathione S-Transferases; LMembrane-Associated Proteins involved in EicosanoidGSTA, GST alpha class; GSTM, GST mu class; GSTP, Gclass; GSTK, GST kappa class; GSTZ, GST zeta class; Gcopy number variants; X2, chi-square test; LD, Linkageancestry in Southwest USA; CEU, Utah residents with Noancestry from the CEPH collection; CHB, Han Chinese inMetropolitan Denver, Colorado; GIH, Gujarati Indians inin Tokyo, Japan; LWK, Luhya in Webuye, Kenya; MEX, MCalifornia; MKK, Maasai in Kinyawa, Kenya; TSI, Tuscan iNigeria; IL-1β, interleukin-1β.⁎ Corresponding author at: Department of Biology, Un

Via della Ricerca Scientifica 1, 00133 Rome, Italy. Fax: +E-mail address: [email protected] (M. Fuciarelli

0378-1119/$ – see front matter © 2012 Elsevier B.V. Alhttp://dx.doi.org/10.1016/j.gene.2012.09.113

a b s t r a c t

Article history:Accepted 29 September 2012Available online 6 October 2012

Keywords:PharmacogeneticsPopulation demographyGenetic polymorphismEcuadorEthiopiaItaly

Glutathione S-Transferase enzymes (GSTs) constitute the principal Phase II superfamilywhich plays a key role incellular detoxification and in other biological processes. Studies of GSTs have revealed that genetic polymor-phisms are present in these enzymes and that some of these are Loss-of-Function (LoF) variants, which affectenzymatic functions and are related to different aspects of human health.The aim of this study was to analyze functional genetic differences in GST enzymes among human populations.Attention was focused on LoF polymorphisms of GSTA1, GSTM1, GSTO1, GSTO2, GSTP1 and GSTT1 genes. TheseLoF variants were analyzed in 668 individuals belonging to six human groups with different ethnic backgrounds:Amhara andOromo fromEthiopia; Colorado and Cayapa Amerindians andAfrican Ecuadorians fromEcuador; andone sample from central Italy. The HapMap database was used to compare our data with reference populationsand to analyze the haplotype and Linkage Disequilibrium diversity in different ethnic groups.Our results highlighted that ethnicity strongly affects the genetic variability of GST enzymes. In particular, GSThaplotypes/variants with functional impact showed significant differences in human populations, according totheir ethnic background. These data underline that human populations have different structures in detoxificationgenes, suggesting that these ethnic differences influence disease risk or response to drugs and therefore haveimplications for genetic association studies involving GST enzymes.In conclusion, our investigation provides data about the distribution of important LoF variants in GST genes inhuman populations. This information may be useful for designing and interpreting genetic association studies.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Genes and environment are included in an “action-reaction” mech-anism, which determines in the expression of the human phenotypes(Li et al., 2012). Genetic predisposition may strongly affect not only aperson's susceptibility to toxic compounds but also their response todrug response (Mroziewicz and Tyndale, 2010). Therefore, the study

oF, Loss-of-Function; MAPEG,and Glutathione metabolism;ST pi class; GSTT, GST thetaSTO, GST omega class; CNVs,Disequilibrium; ASW, Africanrthern and Western EuropeanBeijing, China; CHD, Chinese inHouston, Texas; JPT, Japaneseexican ancestry in Los Angeles,n Italy; YRI, Yoruban in Ibadan,

iversity of Rome “Tor Vergata”,39 062023500.

).

l rights reserved.

of human genetics helps to elucidate many aspects of human health.For example, the study of genes involved in cellular detoxificationmay be useful for analyzing the interaction between genetics and theenvironment (Piacentini et al., 2010). Indeed, detoxification enzymesare directly involved in interactions between living organisms andtheir environments (diet, climate, and lifestyle) (Lampe, 2007). Thecellular detoxification mechanism is broken down into three phases:Phase I (oxidation, reduction and hydrolysis), Phase II (conjugation),and Phase III (excretion) (Omiecinski et al., 2011). The enzymes in-volved in the detoxification processes showed significant inter-ethnicand inter-individual differences in their efficiency (Polimanti et al.,2011a). These differences in the enzymatic systems are due to genetic,and environmental factors andmay explain the ethnic diversity observedin the susceptibility to exposure to some xenobiotic compounds (Thieret al., 2003). In particular, the investigation of ethnic differences in var-iants associated with significant alterations in the coding-proteins,called Loss-of-Function (LoF), may contribute to our understanding atthe population level the genetic predisposition to the disease or todrug response.

Among detoxification enzymes, glutathione S-transferases (GSTs)are multi-functional proteins that constitute the principal superfamily

Table 1Genotype frequencies of GST LoF variants in worldwide populations.

AfricanEcuadoriansn=159

Amharan=100

Cayapasn=114

Coloradosn=78

Italiansn=120

Oromon=97

GSTA1*-69C/TC/C (%) 84 (53) 55 (55) 46 (41) 43 (55) 54 (45) 48 (50)C/T (%) 65 (41) 35 (35) 62 (54) 34 (44) 47 (39) 43 (44)T/T (%) 10 (6) 10 (10) 6 (5) 1 (1) 19 (16) 6 (6)

GSTM1Positive (%) 87 (55) 50 (50) 92 (81) 37 (47) 61 (51) 60 (62)Null (%) 72 (45) 50 (50) 22 (19) 41 (53) 59 (49) 37 (38)

GSTO1*A140DA/A (%) 123 (77) 62 (62) 108 (95) 64 (82) 54 (45) 54 (56)A/D (%) 31 (20) 35 (35) 6 (5) 14 (18) 62 (52) 35 (36)D/D (%) 5 (3) 3 (3) 0 (0) 0 (0) 4 (3) 8 (8)

GSTO1*E155delE/E (%) 148 (93) 95 (95) 98 (86) 69 (88) 109 (91) 94 (97)E/del (%) 11 (7) 5 (5) 16 (14) 9 (12) 11 (9) 3 (3)

GSTO1*E208KE/E (%) 145 (91) 95 (95) 95 (83) 70 (90) 106 (88) 93 (96)E/K (%) 14 (9) 5 (5) 19 (17) 8 (10) 14 (12) 4 (4)

GSTO2*N142DN/N (%) 27 (15) 15 (15) 98 (86) 64 (82) 51 (43) 17 (18)N/D (%) 89 (60) 56 (56) 15 (13) 13 (17) 58 (48) 45 (46)D/D (%) 43 (25) 29 (29) 1 (1) 1 (1) 11 (9) 35 (36)

GSTP1*I105VI/I (%) 55 (35) 64 (64) 45 (40) 32 (41) 63 (53) 60 (62)I/V (%) 66 (41) 29 (29) 41 (36) 27 (35) 47 (39) 31 (32)V/V (%) 38 (24) 7 (7) 28 (24) 19 (24) 10 (8) 6 (6)

GSTT1Positive (%) 143 (90) 61 (61) 111 (97) 68 (87) 86 (72) 63 (65)Null (%) 16 (10) 39 (39) 3 (3) 10 (13) 34 (28) 34 (35)

103R. Polimanti et al. / Gene 512 (2013) 102–107

of Phase II enzymes (Frova, 2006). These enzymes are involved in anumber of catalytic processes, such as reactive electrophiles' detoxi-fication, in biosynthesis of leukotrienes, prostaglandins, testosteroneand progesterone, and tyrosine degradation. These proteins also play arole as non-enzymatic modulatory elements (Hayes et al., 2005). Thisenzymatic superfamily is composed of three different families: mito-chondrial, microsomal (or MAPEG, Membrane-Associated Proteins in-volved in Eicosanoid and Glutathione metabolism), and cytosolic. Thecytosolic family is the most abundant and the cytosolic GSTs are classi-fied in seven classes based on chromosomal location and on sequencesimilarity: alpha (GSTA), mu (GSTM), pi (GSTP), theta (GSTT), kappa(GSTK), zeta (GSTZ) and omega (GSTO). Each cytosolic class is usuallyconstituted by multiple members and is located in a specific chromo-somal cluster. Numerous variants have been identified in GST genes,and some of these polymorphisms can be classified as LoF becausethey are associated with significant alterations in the enzymatic func-tions (Fuciarelli et al., 2009; Josephy, 2010). Several studies haveexplored whether the LoF variants of GSTs are significantly associatedwith disease risk, highlighting positive outcomes for different typesof cancers (Di Pietro et al., 2010) and for other common disorders,such as neurologic (Piacentini et al., 2012a, 2012b), cardio-vascular(Polimanti et al., 2011b), pregnancy-related (Polimanti et al., 2012),and allergic diseases (Piacentini et al., 2012c). These studies pointout that great inter-ethnic diversity is present in the allele frequenciesof some LoF variants of GST genes and that it is possible to observethat ethnicity significantly influences these GST-disease associations(Piacentini et al., 2011). To understand the variability of GST genes inworldwide populations, different studies have analyzed LoF variantsin different ethnic groups, confirming that human demographic historyaffects GST gene distribution (Polimanti et al., 2011c). Unfortunately,most of these studies have focused their attention on the copy numbervariants (CNVs) of GSTM1 and GSTT1 genes, leaving the other LoF GSTvariants poorly investigated (Gaspar et al., 2002; Polimanti et al.,2011a).

The aimof this studywas to analyze functional genetic differences inGST enzymes among human populations. Attention was focused on LoFpolymorphisms of GSTA1 (rs3957357), GSTM1 (CNV), GSTO1 (rs4925,rs11509437, rs11509438), GSTO2 (rs156697), GSTP1 (rs1695) andGSTT1 (CNV) genes that were investigated in populations with African,American and European origins. The selection of variants was based ontheir functional impacts and on their implication in human disease. Toprovide a comprehensive analysis of human diversity, our data werecomparedwith the genetic information available in the HapMap project(International HapMap 3 Consortium et al., 2010). Moreover, HapMapdata were used to analyzed the ethnic differences in the structures ofthe investigated GST genes.

2. Materials and methods

2.1. Subjects

A total of 668 unrelated adult individuals of both sexes have beentyped: Amhara (n=100), Oromo (n=97), Cayapas (n=114), Colorados(n=78), African Ecuadorians (n=159), and Italians (n=120). 5–10 mlof peripheral blood from each subject was collected by venipuncture andstored in heparinized. Each donor was asked to supply name, birthplace,language and ethnicity for three generations, in order to allow us todetermine the extent of recent admixture. Further information aboutthese human groups is available in previous studies (De Angelis et al.,2012; De Stefano et al., 2002; Polimanti et al., 2010).

2.2. Genotyping

Genotyping of rs3957357 (GSTA1*-69C/T), rs4925 (GSTO1*A140D),rs156697 (GSTO2*N142D) and rs1695 (GSTP1*I105V) was performedusing the PCR-restriction fragment length polymorphism (RFLP)method.

rs11509437 (GSTO1*E155del) and rs11509438 (GSTO1*E208K) weretyped using the confronting two-pair primer and the allele-specificmethods, respectively. Genotyping of GSTM1 and GSTT1 CNVs was car-ried out by a Multiplex PCR reaction. Methodologies have beendescribed in our previous studies (Piacentini et al., 2010; Polimantiet al., 2010).

2.3. Statistical analysis

Allele frequencies were computed by the genotype-countingmethod. Hardy–Weinberg equilibrium was evaluated using the chi-square (χ2) test. Population comparisons and AMOVA were performedby Arlequin 3.5.1.2 (Excoffier and Lischer, 2010). Pairwise FST differ-ences and FST P values were calculated to analyze the inter-populationdifferences. One hundred ten permutations of individuals betweenpopulations were computed to test the significance of distances and0.05 was the minimum P value of a test to be considered as significant.To compare theGST allelic frequencies inworldwidepopulations, corre-spondence analysis was utilized (Greenacre, 1992). Linkage disequilib-rium (LD) analysis was performed by Haploview version 3.2 (Barrettet al., 2005) and graphically displayed usingHaploview linkage disequi-librium plots. To identify the functional impact of the analyzed SNPs,FASTSNP (Function Analysis and Selection Tool for SNPs) was used(Yuan et al., 2006).

3. Results

Table 1 shows the genotype frequencies of GST LoF variants observedin the six human groups considered. Genotype distributions were in

104 R. Polimanti et al. / Gene 512 (2013) 102–107

Hardy–Weinberg equilibrium for all loci and populations. To analyze ourresults, we matched our data with those reported for different ethnicgroups in the HapMap project (available at http://hapmap.ncbi.nlm.nih.gov). Our six populations and the eleven HapMap samples can beclassified in four groups considering their geographic origin: Africa(Amhara, Oromo, African Ecuadorians, ASW, LWK, MKK, YRI); Asia(CHB, CHD, JPT); Europe (Italians, CEU, TSI); and America (Cayapas,Colorados, GIH, MEX). Considering this structure, important differencesin the minor allele frequency averages were observed, except forGSTO1*E155del and GSTO1*E208K. For these GSTO1 variants, theminor allele frequencies never exceeded 10% (uncommon variants)and no significant differences were observed among populations withdifferent ethnic backgrounds. Moreover, our results confirm the linkagein the GSTO1 gene between the E155 deletion and E208K substitu-tion in worldwide populations (African Ecuadorians D’=0.807; AmharaD’=0.581; Cayapas D’=1; Colorados D’=1; Italians D’=0.738; OromoD’=0.656). The other GST LoF variants (GSTA1*-69C/T, GSTM1 positive/null genotype, GSTO1*A140D, GSTO2*N142D, GSTP1*I105V, GSTT1positive/null genotype) showed a great range of variability among indi-viduals with different geographic origins: GSTA1*-69 T allele from 10%(Asians) to 42.5% (Europeans), GSTM1 null genotype from 33.4%(Africans) to 54.3% (Europeans), GSTO1*D140 from 14.3% (Americans)to 32% (Europeans), GSTO2*D142 from 25.1% (Americans) to 67.9%(Africans), GSTP1*V105 from 16.3 (Asians) to 54.3 (Americans), andGSTT1 null genotype from 15.2% (America) to 39.2% (Asians). In partic-ular, GST SNPs showed differing percentages of molecular varianceamong the geographic origin groups: 5.6% for GSTA*-69C/T, 10.2% forGSTO1*A140D, 25.0% for GSTO2*N142D and 20.6% for GSTP1*I105V.Using these GST LoF variants, we calculated a matrix of pairwise FSTalong with the FST P values (Table 2). Analyzing the pair-wise FST differ-ences, it is possible to observe that populationswith the same ethnic or-igin have no significant differences within the group. This outcome ismore evident in European and Asian groups than in African and Ameri-can clusters. The correspondence analysis performed on GST allelic fre-quencies confirm that human populations are placed according theirgeographic origin (Fig. 1). Asian populations showed a strong correlationwithin the group and this cluster strictly corresponds to the GSTP1*I105allele and the GSTT1 null genotype. The European group showed thetwo Italian samples (Italians and TSI) close together and the CEU sampleslightly farther. The European group appears to be related with theGSTM1 null genotype and the GSTA1*-69 T and GSTO1*D140 alleles.Regarding the African group, populations are more scattered than thepreviously described groups: YRI and LWK populations are close

Table 2Matrices of pair-wise FST distances. Above the diagonal are FST P values. Below the diaAfroecuadorians; CAY: Cayapas; COL: Colorados).

CEU TSI ITA CHD JPT CHB ASW LWK

CEU .216 .207 .000 .000 .000 .108 .000

TSI .009 .162 .000 .000 .000 .045 .000

ITA .017 .004 .000 .000 .000 .045 .000

CHD .157 .101 .057 .045 .783 .000 .000

JPT .184 .104 .061 .012 .009 .000 .000

CHB .163 .108 .060 .000 .009 .000 .000

ASW .039 .033 .430 .172 .164 .176 .018

LWK .245 .205 .213 .314 .308 .319 .062

MKK .088 .056 .062 .141 .130 .146 .000 .061

YRI .218 .173 .183 .280 .278 .292 .052 .000

AMH .097 .045 .039 .088 .065 .090 .013 .125

ORO .093 .038 .039 .099 .078 .104 .014 .125

AFE .069 .066 .058 .110 .118 .113 .000 .072

MEX .121 .183 .207 .355 .418 .376 .148 .225

GIH .017 .017 .001 .052 .061 .052 .032 .194

CAY .098 .141 .097 .126 .166 .116 .199 .358

COL .092 .118 .075 .075 .125 .071 .183 .340

together, distinct from the others and appear to be correlatedwith the GSTO2*D142 allele; the other African subgroup, which isconstituted by Amhara, Oromo, African Ecuadorians, ASW and MKK,has correspondences with GSTM1 and GSTT1 positive genotypes andGSTA1*-69C and GSTO1*A140 alleles. Regarding American populations,Cayapas, Colorados and GIH seem close to each other with no strongcorrelation with GST alleles, whereas the MEX sample is stronglyrelated to the GSTP1*V105 allele and far from the other populationgroups.

After the single-locus analysis, we analyzed the inter-ethnic differ-ences in the structure of gene regions, in which are located the poly-morphic GST SNPs (GSTA1*-69C/T, GSTO1*A140D, GSTO2*N142D,GSTP1*I05V). In order to identify genetic variants in LD with our con-sidered SNPs, the HapMap LD data (HapMap Data Rel 27 PhaseII+III,Feb09, on NCBI B36 assembly, dbSNP b126) were analyzed. In partic-ular, the SNPs in complete LD with the investigated SNPs in at leastone HapMap population were considered.

Regarding the nucleotide substitution –-69C/T in the GSTA1 gene,13 SNPs were identified in LD with this variant. In Fig. 2, the plotsof LD analyses for the GSTA1 gene in African, Asian and Europeanpopulations are reported. The identification of LD blocks was per-formed according the method of Gabriel et al. (2002). In the Africangroup, three blocks are present and GSTA1*-69C/T (rs3957357) islocated in the 5 kb block. In aside from our variant, only SNPs withlow or no known functions are present. In Asian and Europeangroups, some SNPs present in Africans did not reach minor allele fre-quencies (MAFs) of 5% and therefore they are not analyzed. In bothgroups, a single block of 11 kb is identified. In Asians, this blockincludes, besides our variant, two SNPs located in intronic enhancers(rs6917325, rs10948723), whereas in the European block, onlySNPs with low or no known functions are included. Regarding theaminoacidic substitutions in the GSTO class (GSTO1*A140D andGSTO*N142D), 55 SNPs were identified in LD with these variants.In Fig. 3, the LD plots for African, Asian and European populationsare shown. In the African group, three LD blocks are present. TheGSTO*A140D is located in the 6 kb block, and the GSTO2*N124Din the 12 kb block. In the 6 kb block, besides GSTO1*A140D, otherfunctional variants are present: rs12264196 (regulatory region);rs11191979 (regulatory region); rs7077259 (regulatory region);rs7077729 (regulatory region); rs7906957 (regulatory region);and rs17116754 (regulatory region). In the 12 kb block, on theother hand, where GSTO2*N142D is located, the functional variantsare: rs10509769 (regulatory region); and rs12264844 (regulatory

gonal are pairwise FST differences. (ITA: Italians; AMH: Amhara; ORO: Oromo; AFE:

MKK YRI AMH ORO AFE MEX GIH CAY COL

.000 .000 .000 .000 .000 .153 .072 .000 .000

.000 .000 .000 .000 .000 .000 .000 .000 .000

.000 .000 .000 .000 .000 .009 .234 .000 .000

.000 .000 .000 .000 .000 .000 .000 .000 .000

.000 .000 .000 .000 .000 .000 .000 .000 .000

.000 .000 .000 .000 .000 .000 .000 .000 .000

.802 .072 .207 .207 .423 .027 .054 .000 .000

.000 .667 .000 .000 .000 .000 .000 .000 .000

.108 .018 .036 .000 .000 .000 .000 .000

. 033 .000 .009 .009 .000 .000 .000 .000

.011 .085 .702 .000 .000 .000 .000 .000

.008 .083 .000 .000 .000 .000 .000 .000

.014 .050 .032 .038 .054 .000 .000 .000

.175 .244 .251 .239 .121 .000 .000 .000

.055 .164 .037 .043 .037 .185 .000 .000

.217 .369 .193 .210 .143 .304 .070 .090

.189 .336 .163 .178 .121 .274 .051 .006

Fig. 1. Correspondence analysis by GST allele (empty diamond) frequencies in the analyzed ethnic groups and in the HapMap populations (full circles).

105R. Polimanti et al. / Gene 512 (2013) 102–107

region). In Asians, two blocks are recognized and our GSTO variants areboth located in the 22 kb block. In this block, the other functional vari-ants are: rs11191979 (regulatory region); rs11509439 (missense). In Eu-ropeans, two LD blocks are identified: GSTO1*A140D is located in the4 kb block, and GSTO2*N142D in the 30 kb block. In the GSTO1*A140Dblock, the other functional variant is rs11191979 (regulatory region);whereas in the GSTO2*N142D block, rs11509439 (missense) is found.Regarding I105V substitution in the GSTP1 gene, seven LD SNPs wereidentified. In Fig. 4, the LD plots in African, Asian and European groupsare reported. In Africans, no LD blocks are identified and the I105V vari-ant (rs1695) is in complete LD with a SNP with no known function(rs8191446). In Asians and Europeans, our GSTP1 variants are includedin a LD block of 1 kb. In the Asian group, this block contains twolow-function variants (rs749174 e rs1871042), whereas in Europeansthere is a missense substitution (rs1138272).

Fig. 2. Pair-wise linkage disequilibrium (LDs) are shown, calculated between the GSTA1 SNPvariants in the GSTA1 gene. The intensity of the box shading is proportional to the strengthbox. For boxes without any number, D'=1.

4. Discussion

GSTs are multifunctional proteins involved in several biological pro-cesses and play a key role in cellular detoxification systems (Hayes et al.,2005). Genes encoding for these enzymes show a number of polymor-phic sites associated with significant impact on enzymatic expressionand catalytic activity (Josephy, 2010). A vast literature details theassociation of GST LoF variants with the development and expressionof many common disorders (Bolt and Thier, 2006; Di Pietro et al.,2010; Frova, 2006). Besides the clinical aspects, GST genes presentLoF variants with important diversity in the allelic distributionsamong human populations and these ethnic differences may explainpart of the genetic predisposition to develop a clinical phenotype(Ben Salah et al., 2012; Gaspar et al., 2002; Piacentini et al., 2011;Polimanti et al., 2011c).

s in African, Asian and European populations. The top panel depicts the location of theof the LD (D') for the marker pair, which is also indicated as a percentage within each

Fig. 3. Pair-wise linkage disequilibriums (LDs) are shown, calculated between the GSTO SNPs in African, Asian and European populations. Figure description is reported in Fig. 2.

106 R. Polimanti et al. / Gene 512 (2013) 102–107

The aim of our studywas to analyze the genetic distribution of someLoF variants widely investigated among human populations and toexplore the differences in the genetic structures where these variantsare located.

Our data confirm that ethnicity strongly affects genetic variability ofthe investigatedGST LoF variants, with the exception of GSTO1*E155deland GSTO1*E208K. For these GSTO1 variants, we confirm the previousdata about their low MAFs and high LD in different ethnic groups(Mukherjee et al., 2006; Paiva et al., 2008; Schmuck et al., 2008). Theother investigated polymorphisms showed significant inter-ethnicdifferences that are in accordance with the geographic origin of theinvestigated populations. The Ethiopian groups (Amhara and Oromo)and African Ecuadorian are similar to each other and appear to be relat-ed to the other African populations. The African cluster seems to becorrelated with the GSTO2*D142 allele. This common polymorphismis the most investigated variant of the GSTO2 gene and a number ofstudies have associated it with different clinical phenotypes (Board,2011). Moreover, different studies have hypothesized that this locus isrelated to deficiencies of arsenic and ascorbate metabolisms (Paiva etal., 2010; Zhou et al., 2012). The high frequency of the GSTO2*D142allele in individuals with African origin suggests that these sub-jects may be genetically predisposed to have GSTO2-related clin-ical phenotypes.

Fig. 4. Pair-wise linkage disequilibriums (LDs) are shown, calculated between the GSTP1 S

Cayapa and Colorado Amerindians did not show significant differ-ences within the population, but slight differences present with re-spect to the GIH population and the MEX population. Regarding theGIH population, the difference is probably due to the genetic diversityamong Amerindian populations, especially between populations ofSouth and North America (Hunley and Healy, 2011). Conversely,the particular structure observed in the MEX population is stronglyinfluenced by the sample size.

The Italian sample investigated in our study showed great similar-ity with the Italian sample quoted in the HapMap database (TSI) andno significant difference with the other HapMap population withEuropean origin (CEU) was found. This cluster appears to be relatedto derived alleles of GSTM1, GSTA1 and GSTO1 polymorphisms. TheCNV of GSTM1 is one of the most investigated GST polymorphismsand a vast literature details that this common variantmay be associatedwith various common disorders (Bolt and Thier, 2006). The GSTA1 geneencodes for the most expressed GST isoform in the liver and thisenzyme plays a key role in the detoxification of many toxic compounds(Coles and Kadlubar, 2005). The GSTA1*-69 T allele is associated with asignificant decrease of enzyme expression and different studies havehypothesized that this LoF variants may be associated with hepatic andnon-hepatic diseases (Coles et al., 2001). As with the GSTO2 enzyme,GSTO1 is involved in the metabolism of arsenic and ascorbate and,

NPs in African, Asian and European populations. Figure description is reported in Fig. 2.

107R. Polimanti et al. / Gene 512 (2013) 102–107

furthermore, it plays a role in regulation of proinflammatory mediatorinterleukin-1β (IL-1β) and in biotransformation of α-haloketones(Board, 2011). A140D substitution in GSTO1 has been associated withneurologic, neoplastic and pulmonary disorders (Board, 2011). The pres-ence of high frequencies of these disease-alleles suggests that individualswith European origin have a genetic predisposition to develop diseasesin which GSTA1, GSTM1 and GSTO1 enzymes are involved.

The analysis of GST genes revealed that the Asian populations arevery similar for these genetic markers and, in particular, individualswith Asian origin are carriers of the GSTT1 null genotype. Togetherwith GSTM1 CNV, the deletion polymorphism of the GSTT1 is themost investigated GST polymorphism and a significant body of datahas been accumulated linking this common polymorphism with sev-eral diseases and with increasing susceptibility to different toxiccompounds (Bolt and Thier, 2006). The association of the GSTT1null genotype with Asian origin suggests that these populations aremore associated with GSTT1-related phenotypes.

Our analyses of GST LoF variants revealed significant differences inhuman populations that may be associated with inter-ethnic diversityof health. However, the structure of GST genesmay be strongly differentamong human populations and our LoF variants may be in high LD andinteract with other functional variants. In some cases, our analysis ofLD and predicted functional impact in the HapMap population revealedsignificant functional differences in the genetic structure of GSTs.Regarding GSTA1, no interactions were observed between -69C/Tpolymorphism and other variants with functional impact. Conversely,GSTO polymorphisms (GSTO1*A140D and GSTO2*N142D) showed sig-nificant association with other functional variants. In particular, thesevariant-variant interactions are stronger in populationswith African or-igin than in Asians and Europeans. GSP1*I105V showed an associationwith a functional variant (rs1138272, missense substitution) only inpopulations with European origin.

In conclusion, our study provided an analysis of the human geneticdiversity of some GST variants previously associated with health,suggesting that the observed differences are associated with alteredgenetic predisposition to disease or to toxic susceptibility in someethnic groups. Furthermore, our investigation highlighted that geo-graphic origin groups showed strong LD between the investigatedLoF variants and other functional polymorphisms. These outcomesmay explain the contrasting results observed, in some cases, in dis-ease association analyses and, moreover, may be useful for resultinterpretation and for research design of genetic association studies.

Acknowledgments

The subjects of the investigation were adequately informed about theaims of the study and gave their approval, which is also gratefullyacknowledged. Human studies have been performed in accordance withthe ethical standards as laid down by law. The authors declare that theyhave no conflict of interest. This study was supported by RSA (RicercaScientifica di Ateneo) Grant 2009 from University of Rome “Tor Vergata”and by PRIN 2009–2011 (prot. n. 200975T9EW) fromMIUR.

References

Barrett, J.C., Fry, B., Maller, J., Daly, M.J., 2005. Haploview: analysis and visualization ofLD and haplotype maps. Bioinformatics 21, 263–265.

Ben Salah, G., et al., 2012. Polymorphisms of glutathione S-transferases M1, T1, P1 and A1genes in the Tunisian population: an intra and interethnic comparative approach.Gene 498, 317–322.

Board, P.G., 2011. The omega-class glutathione transferases: structure, function, andgenetics. Drug Metab. Rev. 43, 226–235.

Bolt, H.M., Thier, R., 2006. Relevance of the deletion polymorphisms of the glutathioneS-transferases GSTT1 and GSTM1 in pharmacology and toxicology. Curr. DrugMetab.7, 613–628.

Coles, B.F., Kadlubar, F.F., 2005. Human alpha class glutathione S-transferases: geneticpolymorphism, expression, and susceptibility to disease. Methods Enzymol. 401, 9–42.

Coles, B.F., et al., 2001. Effect of polymorphism in the human glutathione S-transferase A1promoter on hepatic GSTA1 and GSTA2 expression. Pharmacogenetics 11, 663–669.

De Angelis, F., Garzoli, A., Battistini, A., Iorio, A., De Stefano, G.F., 2012. Genetic responseto an environmental pathogenic agent: HLA-DQ and onchocerciasis in northwesternEcuador. Tissue Antigens 79, 123–129.

De Stefano, G.F., et al., 2002. Analysis of three RFLPs of the COL1A2 (Type I Collagen) inthe Amhara and the Oromo of Ethiopia. Ann. Hum. Biol. 29, 432–441.

Di Pietro, G., Magno, L.A., Rios-Santos, F., 2010. Glutathione S-transferases: an overviewin cancer research. Expert Opin. Drug Metab. Toxicol. 6, 153–170.

Excoffier, L., Lischer, H.E., 2010. Arlequin suite ver 3.5: a new series of programs to per-form population genetics analyses under Linux and Windows. Mol. Ecol. Resour.10, 564–567.

Frova, C., 2006. Glutathione transferases in the genomics era: new insights and per-spectives. Biomol. Eng. 23, 149–169.

Fuciarelli, M., Caccuri, A., De Francesca, M., Ferazzoli, F., Piacentini, S., Porreca, F., 2009.Modulation of the GSTT1 activity by the GSTM1 phenotype in a sample of Italianfarm-workers. Arch. Toxicol. 83, 115–120.

Gabriel, S.B., et al., 2002. The structure of haplotype blocks in the human genome.Science 296, 2225–2229.

Gaspar, P.A., et al., 2002. Polymorphisms of CYP1a1, CYP2e1, GSTM1, GSTT1, and TP53genes in Amerindians. Am. J. Phys. Anthropol. 119, 249–256.

Greenacre, M., 1992. Correspondence analysis in medical research. Stat. Methods Med.Res. 1, 97–117.

Hayes, J.D., Flanagan, J.U., Jowsey, I.R., 2005. Glutathione transferases. Ann. Rev. Pharm.Toxicol. 45, 51–88.

Hunley, K., Healy,M., 2011. The impact of founder effects, geneflow, and European admix-ture on native American genetic diversity. Am. J. Phys. Anthropol. 146, 530–538.

International HapMap 3 Consortium, et al., 2010. Integrating common and rare geneticvariation in diverse human populations. Nature 467, 52–58.

Josephy, P.D., 2010. Genetic variations in human glutathione transferase enzymes: signif-icance for pharmacology and toxicology. Hum. Genomics Proteomics 2010, 876940.

Lampe, J.W., 2007. Diet, genetic polymorphisms, detoxification, and health risks. Altern.Ther. Health Med. 13, S108–S111.

Li, M., Peng, R.S., Wei, C., Lu, Q., 2012. A U-Statistic-based random forest approach forgenetic association study. Front. Biosci. 4, 2707–2717.

Mroziewicz, M., Tyndale, R.F., 2010. Pharmacogenetics: a tool for identifying genetic fac-tors in drug dependence and response to treatment. Addict. Sci. Clin. Pract. 5, 17–29.

Mukherjee, B., et al., 2006. Glutathione S-transferase Omega 1 and Omega 2pharmacogenomics. Drug Metab. Dispos. 34, 1237–1246.

Omiecinski, C.J., Vanden Heuvel, J.P., Perdew, G.H., Peters, J.M., 2011. Xenobiotic metab-olism, disposition, and regulation by receptors: from biochemical phenomenon topredictors of major toxicities. Toxicol. Sci. (Suppl. 1), S49–S75.

Paiva, L., Marcos, R., Creus, A., Coggan, M., Oakley, A.J., Board, P.G., 2008. Polymorphism ofglutathione transferase Omega 1 in a population exposed to a high environmentalarsenic burden. Pharmacogenet. Genomics 18, 1–10.

Paiva, L., Hernández, A., Martínez, V., Creus, A., Quinteros, D., Marcos, R., 2010. Associ-ation between GSTO2 polymorphism and the urinary arsenic profile in copperindustry workers. Environ. Res. 110, 463–468.

Piacentini, S., et al., 2010. Glutathione S-transferase gene polymorphisms and air pollu-tion as interactive risk factors for asthma in a multicentre Italian field study: a pre-liminary study. Ann. Hum. Biol. 37, 427–439.

Piacentini, S., Polimanti, R., Porreca, F., Martínez-Labarga, C., De Stefano, G.F., Fuciarelli, M.,2011. GSTT1 and GSTM1 gene polymorphisms in European and African populations.Mol. Biol. Rep. 38, 1225–1230.

Piacentini, S., et al., 2012a. GSTO1*E155del polymorphism associated with increasedrisk for late-onset Alzheimer's disease: association hypothesis for an uncommongenetic variant. Neurosci. Lett. 506, 203–207.

Piacentini, S., et al., 2012b. GSTM1 null genotype as risk factor for late-onset Alzheimer'sdisease in Italian patients. J. Neurol. Sci. 317, 137–140.

Piacentini, S., et al., 2012c. Functional polymorphisms of GSTA1 and GSTO2 genes asso-ciated with asthma in Italian children. Clin. Chem. Lab. Med. 50, 311–315.

Polimanti, R., Piacentini, S., Porreca, F., Fuciarelli, M., 2010. Glutathione S-transferase ωclass (GSTO) polymorphisms in a sample from Rome (Central Italy). Ann. Hum.Biol. 37, 585–592.

Polimanti, R., Piacentini, S., Fuciarelli, M., 2011a. HapMap-based study of human solu-ble glutathione S-transferase enzymes: the role of natural selection in shaping thesingle nucleotide polymorphism diversity of xenobiotic-metabolizing genes.Pharmacogenet. Genomics 21, 665–672.

Polimanti, R., Piacentini, S., Lazzarin, N., Re, M.A., Manfellotto, D., Fuciarelli, M., 2011b.Glutathione S-transferase variants as risk factor for essential hypertension in Italianpatients. Mol. Cell. Biochem. 357, 227–233.

Polimanti, R., Piacentini, S., De Angelis, F., De Stefano, G.F., Fuciarelli, M., 2011c. HumanGST loci as markers of evolutionary forces: GSTO1*E155del and GSTO1*E208Kpolymorphisms may be under natural selection induced by environmental arsenic.Dis. Markers 31, 231–239.

Polimanti, R., et al., 2012. Glutathione S-transferase genes and the risk of recurrent mis-carriage in Italian women. Fertil. Steril. 98, 396–400.

Schmuck, E., et al., 2008. Deletion of Glu155 causes a deficiency of glutathione transfer-ase Omega 1–1 but does not alter sensitivity to arsenic trioxide and other cytotoxicdrugs. Int. J. Biochem. Cell Biol. 40, 2553–2559.

Thier, R., et al., 2003. Markers of genetic susceptibility in human environmental hygieneand toxicology: the role of selected CYP, NAT and GST genes. Int. J. Hyg. Environ.Health 206, 149–171.

Yuan, H.Y., et al., 2006. FASTSNP: an always up-to-date and extendable service for SNPfunction analysis and prioritization. Nucleic Acids Res. 34, W635–W641.

Zhou, H., Brock, J., Liu, D., Board, P.G., Oakley, A.J., 2012. Structural insights into thedehydroascorbate reductase activity of human omega-class glutathione transfer-ases. J. Mol. Biol. 420, 190–203.