www.elsevier.com/locate/braindev

Brain & Development 32 (2010) 105–109

Original article

Genetic variant of glutathione peroxidase 1 in autism

Xue Ming a,*, William G. Johnson b, Edward S. Stenroos b, Audrey Mars c,George H. Lambert d, Steven Buyske e

a Department of Neurosciences and Neurology, UMDNJ-New Jersey Medical School, 90 Bergen Street, DOC 8100, Newark, NJ 07103, USAb Department of Neurology, UMDNJ-Robert Wood Johnson Medical School, Piscataway, NJ, USA

c Hunterdon Medical Center, Hunterdon, NJ, USAd Department of Pediatrics, UMDNJ-Robert Wood Johnson Medical School, New Brunswick, NJ, USA

e Department of Statistics and Biostatistics, Rutgers University, Piscataway, NJ, USA

Received 5 September 2008; received in revised form 1 December 2008; accepted 26 December 2008

Abstract

Genetic factors can contribute to autistic disorder (AD). Abnormal genes of oxidative stress pathways and increased oxidativestress have been reported in autism spectrum disorders. Polymorphisms of genes involved in glutathione metabolism, e.g. GSTP1

and GSTM1 are reportedly associated with autistic disorder. We investigated a GCG repeat polymorphism of a human glutathioneperoxidase (GPX1) polyalanine repeat (ALA5, ALA6 and ALA7) in 103 trios of AD (probands and parents) using the transmissiondisequilibrium test. Significant transmission disequilibrium (p = 0.044) was found in the overall transmission of the three alleles. TheALA6 allele was under transmitted (p = 0.017). These results suggest that possessing this ALA6 allele may be protective for AD.Future study of interaction of the GPX1 GCG repeat and other gene polymorphisms such as the MnSOD ALA16 or the GPX1Pro198Leu polymorphism in this cohort of AD families may shed light in whether the combination of the ALA6 allele with anotherpolymorphism of antioxidant allele contributes to the increased oxidative stress in autism.Published by Elsevier B.V.

Keywords: Glutathione peroxidase 1; Oxidative stress; Autistic disorder; Polymorphism; Transmission/disequilibrium test

1. Introduction

Autism spectrum disorders (ASD) have been associ-ated with alterations of multiple gene variants. Gene-environment interactions may also play a role. Genesof detoxification pathways and oxidative stress havebeen subjects of research in ASD. Several lines of evi-dence support an association of oxidative stress withASD in at least some cases. First, there is evidence ofreduced endogenous antioxidant capacity. Specifically,reduced enzymatic activities of glutathione peroxidase(GPX) [1–3], superoxide dismutase (SOD) [1,3] and cat-

0387-7604/$ - see front matter Published by Elsevier B.V.

doi:10.1016/j.braindev.2008.12.017

* Corresponding author. Tel.: +1 973 972 5204; fax: +1 973 9729553.

E-mail address: mingxu@umdnj.edu (X. Ming).

alase [4], and reduced levels of total glutathione (GSH),GSH/GSSG and cysteine [5] have been reported. Levelsof exogenous antioxidants were also reportedly reducedin autism, including vitamin C, vitamin E and vitamin Ain plasma, and zinc and selenium in erythrocytes [5].

A second indicator of altered oxidative stress in aut-ism is derived from evidence of impaired energy metab-olism. Magnetic resonance spectroscopic study of thebrains of autistic individuals showed reduced synthesisof ATP [6]. In addition, higher lactate [7,8] and pyruvate[9] levels have been reported.

Third, there have been reports of improvement in cer-tain behaviors following antioxidant administration toindividuals with autism. In double-blind, placebo-con-trolled trials, high-dose vitamin C [10] or carnosine[11] improved autistic behavior over baseline observa-

106 X. Ming et al. / Brain & Development 32 (2010) 105–109

tions. Likewise, children with autism, who haddecreased blood levels of the antioxidants GSH and cys-teine as well as a decreased GSH/GSSG ratio comparedwith controls, had increases of these following a 3-weeksupplementation with betaine and folinic acid [5].

Fourth, increased excretion of oxidative stress bio-markers has been reported in children with autism. Spe-cifically, the excretion of a F2 isoprostane, 8-iso-prostaglandin F2a is increased in children with autismspectrum disorders [12]. This isoprostane is a productof nonenzymatic oxidation of arachidonic acid and iswidely recognized as a reliable marker of lipid peroxida-tion [13]. Furthermore, nitric oxide, a free radical thatcan block energy production, was found to be increasedin autism as compared to age and sex-matched controls[2]. In addition, elevated nitrite concentrations havebeen detected in individuals with autism [14] along withelevations of thiobarbituric acid reactive substances andxanthine oxidase activity in red cells [4]. Consistent withthe increased oxidative stress biomarkers, children withASD were found to have increased body burdens ofenvironmental toxins that may generate oxidative stress[15,16]. Taken together, these lines of evidence suggestthat it is likely that at least some children with autismexhibit enhanced oxidative stress. However, none ofthese observations suggest how oxidative stress can leadto autism.

Genes of glutathione metabolism have been impli-cated in ASD. Buyske et al. [17] found that homozygos-ity of the gene deletion polymorphism of glutathioneS-transferase (GST) GSTM1 was associated with aut-ism. Williams et al. [18] reported significant transmissiondisequilibrium of a GSTP1 haplotype in maternal triosof autistic case families. Further, James et al. [5]reported abnormal metabolism in oxidative stresspathways.

GPX is a major enzyme in the glutathione pathwaythat catalyzes the reduction of free radicals by glutathi-one and represents a major enzyme for defense againstoxidant molecules [19]. GPX1, a selenium dependentprotein, is the predominant and most abundant isoen-zyme of GPX. GPX1 is located in cytoplasm and isfound in most parenchymal organs and peripheral bloodcells, particularly erythrocytes. This enzyme is believedto play an integral role in cellular antioxidant defensemechanisms by catalyzing the reduction of potentiallyharmful peroxides, thereby contributing to maintenanceof health by reducing oxidative stress. Altered GPXenzyme activity was reported in spina bifida [20],another neurodevelopmental disorder, idiosyncratic val-proate toxicity [21] and autism [1–3]. Valproate therapyhas been shown to increase serum lipid peroxides and todecrease GPX activity secondary to a valproate-induceddecrease in selenium concentrations. Patients with lowGPX are at risk for the development of idiosyncraticadverse reaction of valproate [21]. Children with spina

bifida had lower erythrocyte GPX activity as comparedwith their healthy age-matched controls. Parents ofthese children also demonstrated lower GPX activity[20]. These observations suggest that oxidative stresscontributes to valproate toxicity and spina bifida in agenetically susceptible populations. GPX levels werefound to be reduced in children with autism [1–3], andfetal valproate syndrome is associated with autism [22–24]. Thus, it is possible that oxidative stress contributesto phenotypes of autism spectrum disorders in the pres-ence of a genetic susceptible variant of GPX.

Polymorphisms of GPX1 were reported to be associ-ated with increased risk of breast cancer (Pro198Leu)[25], and cardiovascular disease (GCG repeat) [26]. Inaddition, mice with a homozygous null mutation forGPX1 show increased susceptibility to the oxidativestress-inducing agents paraquat and hydrogen peroxide[27] and to diquat-induced oxidative stress [28].

The gene for human GPX1 contains a sequence poly-morphism in which a short trinucleotide repeat sequence(GCG) codes for a variable length polyalanine stretchclose to the N-terminus [29]. The polyalanine repeatpolymorphism has three alleles with five (ALA5), six(ALA6) and seven (ALA7) alanine residues [26]. Weinvestigated this common polyalanine repeat polymor-phism of human GPX1 in a cohort of trios of autisticdisorder family (probands and parents). We report hereour finding of disequilibrium in transmission of one ofthe three polymorphic alleles.

2. Materials and methods

2.1. Participants

Families having a child diagnosed with autistic disor-der were invited to participate through advertisementsin the newsletter of the New Jersey Center for Outreachand Services for the Autism Community (COSAC,Ewing, NJ). A few families were recruited through ourDepartment of Pediatrics. Selection criteria for familieswere (1) participation of a proband with the clinical diag-nosis of non-syndromic autistic disorder by their neuro-developmental pediatrician as assessed by telephoneinterview with the primary caregiver, (2) clinical diagnosisof autistic disorder (AD) of 101 probands and pervasivedevelopmental disorder-not otherwise specified (PDD-NOS) of 2 probands confirmed by the Autism DiagnosticInterview-revised (ADI-R) and the Autism DiagnosticObservation Schedule-Generic (ADOS-G) testing[30,31]; (3) both biological parents willing to participate.All probands were tested using both the ADI-R and theADOS-G by a certified examiner (AM). This study wasapproved by the Institutional Review Board of UMDNJand informed consent was obtained.

Venous blood was collected in Vacutainers contain-ing EDTA. The samples were either frozen immediately

X. Ming et al. / Brain & Development 32 (2010) 105–109 107

in cryovials at �70 �C or frozen briefly (2–3 days) at�20 �C for transport to the laboratory and preparedthere. Genomic DNA was obtained from whole bloodor white blood cells isolated by centrifugation. DNAextraction was done using Qiagen QIAamp DNA BloodKits (Qiagen, Valencia, CA). Genomic DNA was storedat 4 �C. The primers used were fluorescent FAM labeled50-/56-FAM/aca att gcg cca tgt gtg-30 and 50-cag cca ctactg cac gt-30. Primers were HPLC purified. PCR condi-tions were as follows: 200 ng template, 200 lM dNTPs,5 pM each primer, 1.5 mM Mg2+ in a 10 ll final volumeand PCR cycle consisted of an initial 94 �C 40 hold, fol-lowed by 94 �C for 5500, 60 �C for 5500, 72 �C for 5500 for35 cycles followed by a final extension at 72 �C for 120.PCR was carried out in a Perkin-Elmer GeneAmp9600 thermal cycler. The PCR product was treated withExoSAP to remove unused primers and dNTPs as permanufacturer’s protocol (USB, Cleveland, Ohio) andproducts were run on an ABI PRISM� 3100 GeneticAnalyzer using GeneScan v3.7 (Foster City, CA). Allelecalls were done using API Genotyper v3.7 NT.

Testing of association of alleles of the GPX1 locus withautistic disorders was done with an extension of the trans-mission/disequilibrium test (TDT) to multi-allelic loci asimplemented in the Mendel software [32]. The TDT com-pares counts of transmitted and non-transmitted allelesfrom parents to offspring. Non-transmitted alleles areused as internal matched control, so the test does notrequire a separate control sample and is protected againstfalse positives caused by population stratification. Men-del’s TDT_Max statistic was calculated as the largestTDT statistic for each allele; the significance of theobserved TDT_Max statistic was then determined by per-mutation testing. Mendel uses case-parent trios with allmembers genotyped as well as duos of a case with one par-ent when those individuals are different heterozygotes.

3. Results

Of the 103 probands, 101 had the clinical diagnosis ofautistic disorder and 2 had the diagnosis PDD-NOS byADI-R and ADOS-G. Males comprised 86% of the pro-bands. The racial ethnic background of the families wasrepresentative of a community-based autistic disorderpopulation. Cognitive levels and comorbidity of the pro-bands were not used as selection criteria.

Table 1Allele-by-allele statisticsa.

GPX1 allele Frequency as reportedin Kote-Jarai et al. [37]

Frequency amongparents

Nt

ALA5 .47 .49 3ALA6 .29 .30 2ALA7 .24 .22 2

a Overall permutation p-value: p = 0.044.

Sixty-eight case trios (probands and parents) weregenotyped for the GPX1 polyalanine repeat polymor-phism; three duos consisting of a case and one parentwith different heterozygotes were also included in theanalysis. Allele frequencies among parents are shownin Table 1. These frequencies are consistent with thosefound in other studies [37]. Since there is no present dataon the functionality of the three polyalanine repeats, thestatistics of the TDT of all three alleles were combinedand a permutation-based p-value was calculated. Trans-mission and non-transmission counts are shown inTable 1, along with allele-by-allele v2 statistics and nom-inal p-values. Permutation testing based on the TDT-MAX statistic showed significant transmission disequi-librium (p = 0.044). Genotypes showed an over repre-sentation of the ALA5 allele under all PCR conditionstested. The ALA6 allele was under transmitted (22transmissions compared to an expected value of 31.5,p = 0.017). Secondary tests of differences in transmissionrates between mothers and fathers and between femaleand male children were not significant.

4. Discussion

This study showed that there was overall transmis-sion disequilibrium of the three ALA alleles in autisticdisorder. An allele for ALA6 of GPX1 was significantlyunder transmitted from parents to probands in thesetrios with autistic disorder, suggesting that possessingthis allele may be protective for autistic disorder. GPXis one of the major enzymes participating in reductionof activated oxygen species through the coupled oxida-tion of reduced glutathione [33]. Reduced GPX functioncould lead to impaired detoxification of xenobiotics, afinding that was reported in autism [15,16]. Further-more, reduced GPX activity was reported in autism[1–3]. However, the limitation of this interpretation isthe limited knowledge of function of the ALA allelesof GPX1 gene. Shen et al. [34], in their study of theALA alleles and GPX1 enzyme activity, showed no con-sistent abnormality of the enzyme activity.

Previous studies of the ALA6 allele reported associa-tion with human diseases. A trend towards lower lungtissue 8OHdG levels in individuals with of ALA6 poly-morphism [35] suggested ALA6 allele may be protectiveagainst oxidative stress, in agreement with our study.

umber of allelesransmitted

Number of allelesnot transmitted

v2 Nominalp-value

9 26 2.60 0.1072 41 5.73 0.0178 22 0.72 0.396

108 X. Ming et al. / Brain & Development 32 (2010) 105–109

GPX1 is located on chromosome 3p. Lung tumorswhich showed 3p chromosomal loss of heterozygosityhad higher levels of 8-OHdG adducts and lower GPX1enzyme activity compared to tumors without 3p loss[35]. However, Winter et al. found that individuals withat least one ALA6 allele exhibit increased risk of coro-nary artery disease [26]. ALA6 and ALA7 were foundto be associated with a second primary tumor in patientswho had squamous cell carcinoma of head and neck[36]. Finally, although Kote-Jarai et al. [37] reportedno association between the risk of young onset prostatecancer and the GCG repeat genotypes, they found a sta-tistically insignificant trend of increased frequency ofGPX1 ALA6/ALA6 genotype in patients with prostatecancer in comparison with controls.

Studies on another GPX1 polymorphism, Pro198-Leu, reported association with human disorders. Ratna-singhe et al. [38], found an increased incidence of theGPX1 Pro198Leu (71% lung cancer patients versus58% for control, p < 0.001) among 315 lung cancer casescompared with matched controls. GPX Pro198Leu wasassociated with an increased risk of bladder cancer [39].An association of the 198Pro/Leu genotype with suscep-tibility to allergic bronchial asthma was found in malesmokers [40]. Analysis of GPX1 (Pro198Leu) allele fre-quency and erythrocyte GPX activity in a Swedish andFinnish population by Forsberg et al. [41] did not revealany significant differences between cases of stroke. How-ever, GPX1 Pro198Leu was associated with loweredGPX1 activity and higher breast cancer risk amongDanish women [42].

GPX1 Pro198Leu was reported to co-segregate withthe human GPX1 ALA6 allele [29]. Hamanishi et al.[43] found that the combination of polymorphisms(ALA6/198Leu) of the GPX1 gene had a 40%decrease in enzyme activity, and this functional vari-ant of GPX1 is associated with increased intima-mediathickness of carotid arteries and risk of cardiovascularand peripheral vascular diseases in type 2 diabeticpatients. In addition, a combination of the polymor-phisms of the GPX1 gene (Pro198Leu) and MnSODgene (Ala16Ala) were associated with an increased riskof breast cancer [44]. These reports raise the questionof whether co-existence of the GPX1 ALA6 allele withthe Pro198Leu allele, or with another antioxidantgenetic variant such as MnSOD (Ala16Ala), mayenhance the significance of the polymorphisms in thiscohort of autistic disorder family. Future studies ofcorrelation of oxidative stress biomarkers, GPX1activity and the genotype of GPX1 could suggest thefunctional significance of the GPX1 ALA6 polymor-phism in autistic disorder. In addition, studies of geneby gene interaction of the GPX1 GCG repeat andGPX1 Pro198Leu polymorphism in this cohort ofAD families may shed light in whether the combina-tion of ALA6 allele with another polymorphism of

GPX1 allele contributes to the reduced GPX activityin autism [1–3].

Acknowledgement

This study was supported by a grant from The CureAutism Now Foundation (Autism Speaks) to Ming X.

References

[1] Golse B, Debray-Ritzen P, Durosay P, Puget K, Michelson AM.Alterations in two enzymes: superoxide dismutase and glutathioneperoxidase in developmental infantile psychosis (infantile autism).Rev Neurol (Paris) 1978;134:699–705.

[2] Sogut S, Zoroglu SS, Ozyurt H, Yilmaz HR, Ozugurlu F, SivasliE. Changes in nitric oxide levels and antioxidant enzyme activitiesmay have a role in the pathophysiological mechanisms involved inautism. Clin Chim Acta 2003;331:111–7.

[3] Yorbik O, Sayal A, Akay C, Akbiyik DI, Sohmen T. Investiga-tion of antioxidant enzymes in children with autistic disorder.Prostaglandins Leukot Essent Fatty Acids 2002;67:341–3.

[4] Zoroglu SS, Armutcu F, Ozen S, Gurel A, Sivasli E, Yetkin O,et al. Increased oxidative stress and altered activities of erythro-cyte free radical scavenging enzymes in autism. Eur ArchPsychiatry Clin Neurosci 2004;254:143–7.

[5] James SJ, Cutler P, Melnyk S, Jernigan S, Janak L, Gaylor DW,et al. Metabolic biomarkers of increased oxidative stress andimpaired methylation capacity in children with autism. Am J ClinNutr 2004;80:1611–7.

[6] Minshew NJ, Goldstein G, Dombrowski SM, Panchalingam K,Pettegrew JW. A preliminary 31P MRS study of autism: evidencefor undersynthesis and increased degradation of brain mem-branes. Biol Psychiatry 1993;33:762–73.

[7] Chugani DC, Sundram BS, Behen M, Lee ML, Moore GJ.Evidence of altered energy metabolism in autistic children. ProgNeuropsychopharmacol Biol Psychiatry 1999;23:635–41.

[8] Coleman M, Blass JP. Autism and lactic acidosis. J Autism DevDisord 1985;15:1–8.

[9] Moreno H, Borjas L, Arrieta A, Saez L, Prassad A, Estevez J,et al. Clinical heterogeneity of the autistic syndrome: a study of 60families. Invest Clin 1992;33:13–31.

[10] Dolske MC, Spollen J, McKay S, Lancashire E, Tolbert L. Apreliminary trial of ascorbic acid as supplemental therapy forautism. Prog Neuropsychopharmacol Biol Psychiatry1993;17:765–74.

[11] Chez MG, Buchanan CP, Aimonovitch MC, Becker M, SchaeferK, Black C, et al. Double-blind, placebo-controlled study of L-carnosine supplementation in children with autistic spectrumdisorders. J Child Neurol 2002;17:833–7.

[12] Ming X, Stein TP, Brimacombe M, Johnson WG, Lambert GH,Wagner GC. Increased excretion of a lipid peroxidation bio-marker in autism. Prostaglandins Leukot Essent Fatty Acids2005;73:379–84.

[13] Montuschi P, Barnes PJ, Roberts 2nd LJ. Isoprostanes: markersand mediators of oxidative stress. FASEB J 2004;18:1791–800.

[14] Zoroglu SS, Yurekli M, Meram I, Sogut S, Tutkun H, Yetkin O,et al. Pathophysiological role of nitric oxide and adrenomedullinin autism. Cell Biochem Funct 2003;21:55–60.

[15] Edelson SB, Cantor DS. The neurotoxic etiology of the autisticspectrum disorder: a replicative study. Toxicol Ind Health2000;16:239–47.

[16] Edelson SB, Cantor DS. Autism: xenobiotic influences. ToxicolInd Health 1998;14:799–811.

X. Ming et al. / Brain & Development 32 (2010) 105–109 109

[17] Buyske S, Williams TA, Mars AE, Stenroos ES, Ming SX, WangR, et al. Analysis of case-parent trios at a locus with adeletion allele: association of GSTM1 with autism. BMC Genet2006;7:8.

[18] Williams TA, Mars AE, Buyske SG, Stenroos ES, Wang R,Factura-Santiago MF, et al. Risk of autistic disorder in affectedoffspring of mothers with a glutathione S-transferase P1 haplo-type. Arch Pediatr Adolesc Med 2007;161:356–61.

[19] Lei XG, Cheng WH, McClung JP. Metabolic regulation andfunction of glutathione peroxidase-1. Annu Rev Nutr2007;27:41–61.

[20] Graf WD, Pippenger CE, Shurtleff DB. Erythrocyte antioxidantenzyme activities in children with myelomeningocele. Dev MedChild Neurol 1995;37:900–5.

[21] Pippenger CE, Meng X, Rothner AD. Free radical scavengingenzyme activity profiles in risk assessment of idiosyncratic drugreactions. Probable mechanism for valproate-induced acutepancreatitis and hepatotoxicity. In: Levy RH, Penry JK,editors. Idiosyncratic reactions to valproate: clinical riskpatterns and mechanisms of toxicity. New York: Raven Press;1991. p. 75–88.

[22] Williams PG, Hersh JH. A male with fetal valproate syndromeand autism. Dev Med Child Neurol 1997;39:632–4.

[23] Christianson AL, Chesler N, Kromberg JGR. Fetal valproatesyndrome: clinical and neurodevelopmental features in two siblingpairs. Dev Med Child Neurol 1994;36:357–69.

[24] Moore SJ, Turnpenny P, Quinn A, Glowver S, Lloyd DJ,Montgomery T, et al. A clinical study of 57 children with fetalanticonvulsant syndrome. J Med Genet 2000;37:489–97.

[25] Knight JA, Onay UV, Wells S, Li H, Shi EJ, Andrulis IL, et al.Genetic variants of GPX1 and SOD2 and breast cancer risk at theOntario site of the Breast Cancer Family Registry. CancerEpidemiol Biomarkers Prev 2004;13:146–9.

[26] Winter JP, Gong YY, Grant PJ, Wild CP. Glutathione peroxidase1 genotype is associated with an increased risk of coronary arterydisease. Coron Artery Dis 2003;14:149–53.

[27] De Haan JB, Bladier CC, Griffiths P, Kelner M, O’Shea RD,Cheung NS, et al. Mice with a homozygous null mutation for themost abundant glutathione peroxidase, Gpx1, show increasedsusceptibility to the oxidative stress-inducing agents paraquat andhydrogen peroxide. J Biol Chem 1998;273:22528–36.

[28] Fu Y, Cheng WH, Porres JM, Ross DA, Lei XG. Knockout ofcellular glutathione peroxidase gene renders mice susceptible todiquat-induced oxidative stress. Free Radic Biol Med1999;27:605–11.

[29] Moscow JA, Schmidt L, Ingram DT, Gnarra J, Jonson B, CowanKH. Loss of heterozygosity of the human cytosolic glutathioneperoxidase 1 gene in lung cancer. Carcinogenesis1994;15:2769–73.

[30] Lord C, Rutter M, Le Couteur A. Autism Diagnostic Interview-Revised: a revised version of a diagnostic interview for caregiversof individuals with possible pervasive developmental disorders. JAutism Dev Disord 1994;24:659–85.

[31] Lord C, Risi S, Lambrecht L, Cook Jr EH, Leventhal BL,DiLavore PC, et al. The autism diagnostic observation schedule-generic: a standard measure of social and communication deficits

associated with the spectrum of autism. J Autism Dev Disord2000;30:205–23.

[32] Lange K, Cantor R, Horvath S, Perola M, Sabatti C, SinsheimerJ, et al. Version 4.0: a complete package for the exact geneticanalysis of discrete traits in pedigree and population data sets. AmJ Hum Genet 2001;69:A1886.

[33] Flohe L. The selenoprotein glutathione peroxidase. In: DolphinD, Avramovic O, Poulson R, editors. Glutathione: chemical,biochemical and medical aspects. New York: John Wiley; 1989. p.644–731.

[34] Shen Q, Townes PL, Padden C, Newburger PE. An in-frametrinucleotide repeat in the coding region of the human cellularglutathione peroxidase (GPX1) gene: in vivo polymorphism andin vitro instability. Genomics 1994;23:292–4.

[35] Hardie LJ, Briggs JA, Davidson LA, Allan JA, King RFGJ,Williams GI, et al. Glutathione peroxidase 1 genotype isassociated with an increased risk of coronary artery disease.Carcinogenesis 2000;21:167–72.

[36] Jefferies S, Kote-Jarai Z, Goldgar D, Houlston R, Frazer-Williams MJ, A’Hern R, et al. Association between polymor-phisms of the GPX1 gene and second primary tumours after indexsquamous cell cancer of the head and neck. Oral Oncol2005;41:455–61.

[37] Kote-Jarai Z, Durocher F, Edwards SM, Hamoudi R, JacksonRA, Ardern-Jones A, et al. Association between the GCGpolymorphism of the selenium dependent GPX1 gene and therisk of young onset prostate cancer. Prostate Cancer Prostatic Dis2002;5:189–92.

[38] Ratnasinghe D, Tangrea JA, Andersen MR, Barrett MJ, VirtamoJ, Taylor PR, et al. Glutathione peroxidase codon 198 polymor-phism variant increases lung cancer risk. Cancer Res2000;60:6381–3.

[39] Ichimura Y, Habuchi T, Tsuchiva N, Wang L, Oyama C, Sato K,et al. Increased risk of bladder cancer associated with aglutathione peroxidase 1 codon 198 variant. J Urol2004;172:728–32.

[40] Solodilova MA, Ivanov VP, Polonikov AV, Khoroshaia IV,Kozhukhov MA, Panfilov VI. Heterozygosity of 198LEU mutantallele in glutathione peroxidase-1 gene as a risk factor of bronchialasthma associated with smoking. Ter Arkh 2007;79:33–6.

[41] Forsberg L, de Faire U, Morgenstern R. Oxidative stress, humangenetic variation, and disease. Arch Biochem Biophys2001;389:84–93.

[42] Ravn-Haren G, Olsen A, Tjonneland A, Dragsted LO, Nexo BA,Wallin H, et al. Associations between GPX1 Pro198Leu poly-morphism, erythrocyte GPX activity, alcohol consumption andbreast cancer risk in a prospective cohort study. Carcinogenesis2006;27:820–5.

[43] Hamanishi T, Furuta H, Kato H, Doi A, Tamai M, ShimomuraH, et al. Functional variants in the glutathione peroxidase-1(GPx-1) gene are associated with increased intima-media thick-ness of carotid arteries and risk of macrovascular diseases inJapanese type 2 diabetic patients. Diabetes 2004;53:2455–60.

[44] Cox DG, Tamimi RM, Hunter DJ. Gene x Gene interactionbetween MnSOD and GPX-1 and breast cancer risk: a nestedcase-control study. BMC Cancer 2006;31:217.