Estrogen-Metabolizing Gene Polymorphisms and Lipid Levels in Women with Different Hormonal Status

Estrogen-metabolizing gene polymorphisms and lipid levels in women with different hormonal status

S Almeida1 ABSTRACT ´1 Endogenous and exogenous sex steroid hormones have multiple effects on MR Zandona lipid and lipoprotein metabolism. It is also known that estrogen has 1 N Franken antiatherogenic actions, therefore we considered examining whether there SM Callegari-Jacques2 was any association between polymorphisms in estrogen-metabolizing genes MC Oso´rio-Wender3 and lipid levels in women. We investigated the association between variants 1 in genes related to estrogen biosynthesis (CYP19-TTTAn) and estrogen MH Hutz catabolism (CYP1A1*2A, CYP1A1*2C, CYP1A2-Asn516Asn, CYP3A4*1B, and 1Departamento de Gene´tica, Instituto de COMT-Val158Met) with serum lipid levels in a cross-sectional study with 472 Biocieˆncias, Universidade Federal do Rio Grande Brazilian women of European descent. They were divided into three do Sul, Porto Alegre, RS, Brazil; 2Departamento subgroups according to their hormonal status: premenopausal women de Estatı´stica, Instituto de Matema´tica, (n ¼ 187), postmenopausal women exposed to hormonal replacement Universidade Federal do Rio Grande do Sul, Porto Alegre, RS, Brazil; 3Departamento de Ginecologia therapy (HRT) (n ¼ 118), and postmenopausal women unexposed to HRT e Obstetrı´cia, Hospital de Clı´nicas de Porto (n ¼ 167). The postmenopausal women receiving HRT who were carriers of Alegre, Universidade Federal do Rio Grande do the CYP3A4*1B variant showed lower low-density lipoprotein cholesterol Sul, Porto Alegre, RS, Brazil levels than wild-type homozygotes. Premenopausal women homozygous for the CYP1A1*2C allele had higher high-density lipoprotein cholesterol levels Correspondence: Professor MH Hutz, Departamento de than heterozygotes. While the CYP1A1*2C variant probably has a higher Gene´tica, Instituto de Biocieˆncias, catalytic activity, the functional implications of the CYP3A4 polymorphism Universidade Federal are still uncertain. These data are the first attempt to associate estrogen do Rio Grande do Sul, Caixa Postal 15053, metabolism genes to lipid levels in women. 91501-970 Porto Alegre, RS, Brazil. Tel: þ 55 51 3316 6720 The Pharmacogenomics Journal (2005) 5, 346–351. doi:10.1038/sj.tpj.6500329; Fax: þ 55 51 3343 5850 published online 30 August 2005 E-mail: [email protected] Keywords: estrogen; lipid levels; estrogen-metabolizing gene polymorphisms; hormonal replacement therapy; pharmacogenetics

INTRODUCTION The oral estrogen/hormone replacement therapy (ERT/HRT) in postmenopausal women decreases serum total cholesterol (T-chol) and low-density lipoprotein cholesterol (LDL-C) concentrations, as well as increases serum high-density lipoprotein cholesterol (HDL-C) and triglyceride (TG) levels.1 However, the effects of ERT/HRT may be variable according to metabolic individuality of women;2,3 thus, it is important to identify factors, such as genetic variability, that may alter estrogen concentrations. Genetic polymorphisms play a critical role in determining variation in the expression of several enzymes involved in estrogen Received: 4 April 2005 metabolism.4 Revised: 30 June 2005 Accepted: 19 July 2005 After the menopause, estrogen biosynthesis from androgens and androgen Published online: 30 August 2005 precursors by aromatase in extra-gonadal tissues is the main source of circulating Estrogen-metabolizing gene polymorphisms and lipid levels S Almeida et al 347

estrogens.5 Aromatase is the product of the CYP19 gene, as two length variants, soluble (S-COMT) and membrane- which is a member of the cytochrome P450 (CYP450) bound (MB-COMT),21 which are encoded by a single gene superfamily of genes. It has a tetranucleotide repeat localized to chromosome 22q11.1–2.22 A single base pair polymorphism (TTTA)n ¼ 7À13 in intron 4 and a 3 bp deletion change (G4A) in exon 4 of the COMT gene, in the position 50 bp upstream of the repeat. This deletion is found in those 472 in the long mRNA, and in the 322 in the short mRNA, with seven repeats, generating two alleles: seven repeats results in an amino-acid change (Val4Met), in codon 158 of with the 3 bp deletion [7r(À3)] and seven repeats without MB-COMT and in codon 108 of S-COMT, that decreases the the deletion (7r).6,7 The tetranucleotide repeat polymorph- activity level of the COMT enzyme 3–4-fold.23 ism in intron 4 is mapped about 80 bp downstream of exon Among the genes involved in estrogen metabolism, some 4.6,7 The different genetic variants may lead to alternate may have an interesting potential to evaluate the clinical splicing patterns and mRNA transcripts, which could response to HRT because they (1) are responsible for the last ultimately result in a protein with modified activity. estrogen biosynthesis step (CYP19), (2) have high catalytic Although no data exist on the functionality of either activity in the liver for the formation of less active CYP19 polymorphism, they have been associated with metabolites (CYP1A2 and CYP3A4), (3) have high catalytic estrogen concentrations.8 activity in extra-hepatic tissues for the formation of more The phase I or estrogen oxidative catabolism is mediated active metabolites (CYP1A1), or (4) are responsible by by CYP450 enzymes, which catalyze NADPH-dependent conjugation of catechol estrogens (COMT). Nevertheless, monooxygenation of this hormone to yield more polar these genes have never been investigated in relation to derivatives.9 The principal CYP isoforms responsible for serum lipid levels in women. Therefore, the purpose of this hepatic metabolism of 17b-estradiol (E2) and estrone (E1) are study was to access the possible association of polymorphic the CYP1A2 and CYP3A4, which are mainly involved in variation in these genes in women with different hormonal estrogen 2- and 4-hydroxylations; the CYP1A1 catalyzes status on lipid and lipoprotein levels. estrogen 2-, 15a-, 5a-, and 4-hydroxylations principally in extra-hepatic tissues.10 Catechol estrogens (2-hydroxyestradiol/E1 and 4-hydroxyestradiol/E1) are major metabolites of RESULTS estrogens in humans, and their inactivation is catalyzed by The genotype frequencies observed for all studied poly- catechol-O-methyltransferase (COMT), a conjugating en- morphisms did not reveal statistically significant differences 11 zyme of phase II. It is possible that individual variations compared to those expected under Hardy–Weinberg equili- in metabolic activities in each phase or in the coordination brium. The allele frequencies for all genes in the studied 12 of these two phases regulate the clearance of estrogens. women are presented in Table 1. Thus, in pharmacogenetic studies, genotyping of poly- The women were divided into three groups according to morphic alleles encoding drug-metabolizing enzymes can CYP19 genotypes:24 I—combinations of the shortest alleles be useful for identification of drug metabolism phenotypes.4 Several polymorphisms have been described in the CYP1A1 gene. Two single nucleotide polymorphisms (SNPs) 0 Table 1 Allele frequencies of the polymorphisms investigated have been extensively studied, one in the 3 -non-coding in the total sample region, 3801T4C denominated CYP1A1*2A according to the recommended nomenclature for the genetic poly- Polymorphisms Alleles Frequency morphisms in human P450 genes13,14 and the other within exon 7, CYP1A1*2C.15 These polymorphisms are genetically CYP19 % (SE)a Nb linked and provide at least three-fold increases in the gene’s 7(À3) 39.4 (1.59) 186 catalytic activity.16 7 20.4 (1.31) 96 The CYP3A family and CYP1A2 are the major forms of 8 8.1 (0.89) 38 CYP450 enzymes in the human liver, accounting for 9 0.2 (0.11) 1 approximately 30 and 13% of total CYP protein content, 10 0.1 (0.10) 0 11 28.4 (1.47) 134 respectively.17 The CYP3A4*1B variant is an A4G alteration 12 2.3 (0.49) 11 that occurs in the nifedipine-specific element, in the 50 18 13 1.1 (0.34) 5 regulatory region of the CYP3A4 gene. The effect of the CYP1A1 CYP3A4*1B allele on CYP3A4 activity remains controversial. 2A 22.0 (1.33) 104 Clinical studies have postulated that the CYP3A4*1B variant 2C 14.0 (1.11) 66 results in decreased enzymatic activity; microsomal studies CYP1A2 could not confirm an effect of this variant on enzymatic T 50.0 (1.62) 236 function, whereas a higher CYP3A4 expression of the CYP3A4 CYP3A4*1B allele was reported in vitro.19 AC4T change in 1B 8.0 (0.86) 38 the 295 position of the CYP1A2 gene (rs2470890) was COMT Met 45.0 (1.61) 212 described20; this variant is a silent mutation in codon 516 (Asn516Asn). The functional implication of this variant has aStandard error. not been investigated. In humans, the COMT protein exists bNumber of individuals.

www.nature.com/tpj Estrogen-metabolizing gene polymorphisms and lipid levels S Almeida et al 348

Table 2 High-density lipoprotein means according to the CYP19, CYP1A1*2C, and CYP1A2 genotypes in premenopausal women, and postmenopausal women HRT users, and HRT nonusers

Genotypes P pC

Mean7SD n Mean7SD n Mean7SD n

CYP19 I II III Premenopausal 1.2070.28 97 1.2070.27 67 1.2770.35 16 0.629 0.943 HRT+ 1.7070.37 41 1.5170.32 61 1.5670.37 14 0.035 0.210 HRTÀ 1.2570.35 76 1.2470.28 73 1.2070.24 16 0.816 0.979

CYP1A1*2C 2C/2C 1A/2C 1A/1A Premenopausal 1.4670.28a 9 1.1470.25b 55 1.2070.28a,b 119 0.004 0.024 HRT+ — — 1.5270.29 22 1.6070.36 95 0.414 0.828 HRTÀ — — 1.2470.28 29 1.2470.31 135 0.905 0.905

CYP1A2 C/C T/C T/T Premenopausal 1.1970.27 63 1.1970.31 74 1.2170.23 45 0.945 1.000 HRT+ 1.5570.33 30 1.6270.39 53 1.5470.30 35 0.383 1.000 HRTÀ 1.1370.27 34 1.2470.29 86 1.3170.34 45 0.024 0.144

Values in mmol/l. HRT+, postmenopausal women on hormonal replacement therapy; HRTÀ, postmenopausal women HRT nonusers; pC, corrected P. a,bMultiple comparisons Tukey HSD test 2C/2C and 1A/2C P ¼ 0.002, pC ¼ 0.012; 2C/2C and 1A/1A, P ¼ 0.014, pC ¼ 0.084.

Table 3 Low-density lipoprotein levels according to the For the CYP3A4 gene polymorphism, the statistical CYP3A4 genotype in premenopausal women, and postmeno- analyses were performed only in homozygotes for the pausal women HRT users, and HRT nonusers wild-type allele and heterozygotes for the variant allele, because only four homozygotes for this variant were Genotypes P pC detected in all samples. Only the postmenopausal women HRT þ carriers of the CYP3A4*1B variant showed lower 1A/1A 1A/1B LDL-C levels than wild-type homozygotes (2.8170.37 and 3.4070.79 mmol/l, respectively; P ¼ 0.0005); this associa- Mean7SD n Mean7SD n tion remained significant after correction for multiple testing (pC ¼ 0.003; Table 3). Premenopausal 3.2070.78 142 3.0970.78 35 0.523 1.000 HRT+ 3.4070.79 107 2.8170.37 10 0.0005 0.003 No significant effect was detected for CYP1A1*2A and HRTÀ 4.0071.02 147 3.8670.93 14 0.625 1.000 Val158Met (COMT) polymorphisms on lipid levels in the three groups of women (data not shown). Values in mmol/l. HRT+, postmenopausal women on hormonal replacement therapy; HRTÀ, postmenopausal women HRT nonusers; pC, corrected P.

DISCUSSION The allele and genotype frequencies observed for the C4T (7(À3)/7(À3); 7(À3)/7; 7(À3)/8; 7/7; 7/8; 8/8; 8/9; 8/10); polymorphism in CYP1A2 and the Val158Met polymorph- II—combinations of the shortest and longest alleles (7(À3)/ ism in the COMT gene were similar to those reported in 11; 7(À3)/12; 7(À3)/13; 7/11; 7/12; 7/13; 8/11); and III— other European or European-derived populations.25,26 The combinations of the longest alleles (11/11; 11/12; 11/13). allele frequencies of the CYP1A1*2A, CYP1A1*2C, and The combination I of CYP19 polymorphism and CYP1A2*T CYP3A4*1B variants were similar to those previously allele was associated with higher HDL-C levels in HRT þ and described for the Brazilian population of European des- HRTÀ postmenopausal women, respectively, but after multi- cent4,27,28 and slightly higher than those observed in ple testing, correction P-values did not remain statistically European populations.4,29,30 The frequencies of the CYP19 significant (Table 2). Premenopausal women homozygous tetranucleotide polymorphism were quite similar to those for the CYP1A1*2C allele had higher HDL-C levels than previously detected in another sample from the same heterozygotes (1.4670.28 and 1.1470.25 mmol/l, respec- population described herein.31 tively; P ¼ 0.004); this effect was still detected after multiple The effect of ERT/HRT on lipid levels or atherosclerosis testing correction (pC ¼ 0.024; Table 2). This genotype was progression in women may differ according to genotypes of not observed in postmenopausal women, and therefore its estrogen receptor genes (ESR1 and ESR2).32–34 Since estrogen effect on HRT users was not tested. does have an antiatherogenic action, associations between

The Pharmacogenomics Journal Estrogen-metabolizing gene polymorphisms and lipid levels S Almeida et al 349

the effect of ERT/HRT on lipid levels and genetic variations The results reported represent an initial effort to under- in genes involved in estrogen metabolism are expected, stand the genetic determinants of lipid levels in women in because the products of these genes regulate estrogen the presence of endogenous or exogenous estrogens. concentrations. Although it is known that several of the Although the lipoprotein levels are only intermediate P450 enzymes are responsible for critical steps in the biomarkers for cardiovascular disease, our results suggest conversion of E1 to estradiol and its subsequent catabolism, that the CYP3A4*1B and CYP1A1*2C genotypes may to the best of our knowledge this is the first study to modulate estrogen effects on cardiovascular risk in women. investigate the association between CYP19, CYP1A1, This information may help elucidate pathways through CYP1A2, CYP3A4, and COMT polymorphisms with circulat- which variations in these genes may influence coronary ing lipid levels in women exposed and unexposed to heart disease in postmenopausal women. Prospective fol- endogenous or exogenous estrogen. low-up investigations are clearly needed to elucidate if and Two well-studied polymorphisms genetically linked in the how these polymorphisms affect safety and efficacy of HRT CYP1A1 gene, CYP1A1*2A and CYP1A1*2C, provide at least and if there is a subset of women that would benefit more three-fold increases in its catalytic activity.16 Estrogens are from TRH to prevent cardiovascular disease. Since the data quickly metabolized, and therefore estrogen levels decrease presented here are the first attempt to associate estrogen and estrogen metabolite levels increase in carriers of the metabolism genes to lipid levels, replications of the present CYP1A1*2C variant. The higher HDL-C levels in premeno- findings in larger samples are warranted. pausal women homozygous for the variant CYP1A1*2C were an unexpected finding, because the correlation between HDL-C levels and estrogens is positive, but it may be METHODS explained by one of the estrogen metabolites to CYP1A1 Study Subjects hydroxylation, the 4-hydroxyestradiol/E1, that retains po- The sample consisted of 472 Brazilian women of European tent hormonal activity and a higher association rate with descent. These women went to two clinical centers at the estrogen receptor than estradiol.10,35 Universidade Federal do Rio Grande do Sul for routine blood It is known that the CYP3A family accounts for approxi- tests or a gynecological clinical visit. In this cross-sectional mately 30% of the total hepatic CYP content.17 Levels of study, women were divided into three subgroups according CYP3A4 are approximately 4- to 14-fold higher than those of to hormonal status. The sample characteristics and demo- CYP1A2. This abundance defines CYP3A4 as the major CYP graphic variables were described elsewhere.32 Briefly, in the isoform responsible for the hepatic metabolism of estro- sample considered, 187 were premenopausal women (mean gens.9,17 The present results demonstrated an effect of the age ¼ 3479.7 years, 18–50 years), 118 were postmenopausal CYP3A4*1B variant only in exogenous estrogen presence. women receiving oral HRT with estrogen or estrogen plus We hypothesize that some differences in exogenous estro- progestagen (0.625 mg of conjugated equine estrogen, gen molecules might confer modifications in metabolism n ¼ 24; 0.625 mg of conjugated equine estrogen plus 2.5 mg when compared to endogenous estrogens. Bennink36 re- of medroxyprogesterone acetate, n ¼ 62; 2 mg of estradiol ported that ethinyl estradiol is more potent than estradiol, plus noretisterone acetate, n ¼ 32) for at least 4 months (HRT because the ethinyl substitution in the C17 position inhibits users—mean age ¼ 5676.7 years, range 39–75 years), and the first-pass hepatic metabolism. 167 were postmenopausal women who were not receiving There were some limitations to our study. It might any HRT (HRT nonusers—mean age ¼ 5879.8 years, range have been exposed to selection bias and factors such as 38–85 years) for at least 1 year. The women were considered population stratification, socioeconomic status, and/or postmenopausal if they had no vaginal bleeding for at least health care, which may act in different ways in HRT users 12 months without other obvious pathological or physiolo- and nonusers. In addition, the CYP1A2 and CYP3A4 activity gical cause or bilateral oophorectomy. The premenopausal can be influenced by a number of factors including body women had regular menstrual bleeding within the previous size, tobacco smoke, coffee intake, cruciferous vegetables, 3 months. Information about health and lifestyle factors and some medications.37 The observed associations might (physical activity, smoking status, alcohol consumption and be spurious (type I error) due to multiple statistical hormonal contraceptives, or drugs intake) was obtained comparisons, but we performed the Benjamini and from each individual by an interview. After answering the Hochberg False Discovery Rate procedure for multiple interview, women, wearing light clothes, had their body testing corrections. However, when the probability of the weight and height recorded and their body mass index type I error (false positive) decreases, the probability of (BMI) calculated by the ratio of weight (in kg) and height the type II error (false negative) increases. The appropriate squared (in meters). Exclusion criteria were secondary application of the multiple testing corrections is not always hyperlipidemia due to renal, hepatic or thyroid disease, clear38 and therefore the two P-values (before and after and diabetes or fasting blood glucose levels above 6.9 mmol/ corrections) are shown. Although the possibility of statis- l.39 Individuals who were taking lipid-lowering medication tical error could not be excluded, this work may be were excluded. Informed consent to the blood samples considered as exploratory and suggest candidate poly- drawn for DNA extraction to be used in studies approved by morphisms for association with lipid levels and for HRT the University Ethics Committee was obtained from each pharmacogenetics studies. subject included in the sample.

www.nature.com/tpj Estrogen-metabolizing gene polymorphisms and lipid levels S Almeida et al 350

Measurements of Lipid Serum Levels ACKNOWLEDGEMENTS Blood samples were collected after 12 h fasting. T-chol, HDL- The financial support was provided by Programa de Apoio a Nu´cleos C, TG, and glucose levels were determined by standard de Exceleˆncia (PRONEX, Brazil) and Conselho Nacional de Desen- methods using commercial kits. LDL-C was calculated volvimento Cientı´fico e Tecnolo´gico (CNPq, Brazil). according to the Friedewald formula.40 DUALITY OF INTEREST None declared. DNA Analyses Genomic DNA was extracted from peripheral blood leuko- ABBREVIATIONS cytes by a salting-out procedure.41 The polymerase chain ANOVA one-way analysis of variance reaction (PCR) for the analysis of a C4T polymorphism in COMT catechol-O-methyltransferase the CYP1A2 gene was carried out in a total volume of 25 ml CYP450 cytochrome P450 containing 0.5–1.0 mg genomic DNA, 10 pmol each of the CYP19 cytochrome P450, family 19 upstream and downstream primers (CYP1A2*11F—50ATC CYP1A1 cytochrome P450, family 1, subfamily A, polypeptide 1 0 0 CYP1A2 cytochrome P450, family 1, subfamily A, polypeptide 2 TAC GGG CTG ACC ATG AA3 and CYP1A2*11R—5 TTG CYP3A4 cytochrome P450, family 3, subfamily A, polypeptide 4 0 AGA TGG GGT CTT GCT CT3 ), 200 mmol each of four dNTP, E1 estrone b 1.5 mM of MgCl2, 0.5 U Taq polymerase and 2.5 ml of PCR E2 17 -estradiol buffer. After an initial denaturation at 941C for 5 min, ERT estrogen replacement therapy ESR1 estrogen receptor 1 followed by 35 cycles at 941C for 30 s, 551C for 30 s, 721C for ESR2 estrogen receptor 2 30 s, and a final extension at 721C for 5 min, the PCR HDL-C high-density lipoprotein cholesterol products were subjected to restriction enzyme digestion at HRT hormone replacement therapy 371C for 2 h with MfeI. To visualize CYP1A2 genotypes, LDL-C low-density lipoprotein cholesterol SD standard deviation digestion products were run on 2% agarose gel with SNPs single nucleotide polymorphisms ethidium bromide (336, 200, and 136 bp). The SNPs T-chol total cholesterol CYP1A1*2A, CYP1A1*2C, CYP3A4*1B, and Val4Met varia- TG triglyceride tion in COMT were amplified by PCR, using the same conditions and oligonucleotide primers as previously described.42–44 The amplification products were subsequently digested with the following restriction enzymes under REFERENCES 1 Godsland IF. Effects of postmenopausal hormone replacement conditions recommended by the manufacturer: MspI(CY- therapy on lipid, lipoprotein, and apolipoprotein (a) concentrations: P1A1*2A), BsrDI(CYP1A1*2C), PstI(CYP3A4*1B), and analysis of studies published from 1974–2000. Fertil Steril 2001; 75: Hsp92II (COMT). The genotypes were determined after 898–915. electrophoresis on agarose or acrylamide gels stained with 2 Herrington DM, Klein KP. Invited review: pharmacogenetics of estrogen replacement therapy. J Appl Physiol 2001; 91: 2776–2784. ethidium bromide, using a 100 bp ladder to score the band 3 Tempfer CB, Riener EK, Hefler LA, Huber JC, Muendlein A. DNA sizes. The TTTAn and 3 bp deletion polymorphisms in the microarray-based analysis of single nucleotide polymorphisms may be CYP19 gene were analyzed as described by Arvanitis et al.45 useful for assessing the risks and benefits of hormone therapy. Fertil Steril 2004; 82: 132–137. 4 Cavalli SA, Hirata MH, Hirata RD. Detection of MboII polymorphism at the 50 promoter region of CYP3A4. Clin Chem 2001; 47: 348–351. Statistical Analyses 5 Neven P. The origin of postmenopausal oestrogens. Eur J Cancer 2002; 38: S29–S30. Allele frequencies were estimated by gene counting. The 6 Kurosaki K, Saitoh H, Oota H, Watanabe Y, Kiuchi M, Ueda S. Combined agreement of genotype frequencies with Hardy–Weinberg polymorphism associated with a 3-bp deletion in the 50-flanking region expectations was tested using the w2 test. All other tests and of a tetrameric short tandem repeat at the CYP19 locus. Nippon Hoigaku transformations were performed with the SPSS 8.0 statistical Zasshi 1997; 51: 191–195. 7 Polymeropoulos MH, Xiao H, Rath DS, Merril CR. Tetranucleotide repeat package. These analyses were carried out separately for polymorphism at the human prostatic acid phosphatase (ACPP) gene. premenopausal, postmenopausal women HRT þ and HRTÀ. Nucleic Acids Res 1991; 19: 4792. For association analyses, the lipid and lipoproteins levels 8 Tworoger SS, Chubak J, Aiello EJ, Ulrich CM, Atkinson C, Potter JD et al. were adjusted for age, BMI, and oral contraceptive usage (in Association of CYP17, CYP19, CYPB1, and COMT polymorphisms with serum and urinary sex hormone concentrations in postmenopausal premenopausal women) by multiple regression analyses. TG women. Cancer Epidemiol Biomarkers Prevent 2004; 13: 94–101. levels were ln-transformed to remove skewness and its 9 Badawi AF, Cavalieri EL, Rogan EG. Role of human cytochrome P450 geometric means are shown in the tables. Adjusted means 1A1, 1A2, 1B1, and 3A4 in the 2-, 4-, and 16alpha-hydroxylation of of lipoprotein and lipid levels were compared among 17beta-estradiol. Metabolism 2001; 50: 1001–1003. 10 Lee AJ, Cai MX, Thomas PE, Conney AH, Zhu BT. Characterization of the genotypes of the polymorphisms by one-way analysis of oxidative metabolites of 17beta-estradiol and estrone formed by 15 variance (ANOVA) or Student’s t-test. When the ANOVA selectively expressed human cytochrome p450 isoforms. Endocrinology provided significant results, the Tukey post hoc test for 2003; 144: 3382–3398. multiple comparisons was used. The Benjamini and Hoch- 11 Matsui A, Ikeda T, Enomoto K, Nakashima H, Omae K, Watanabe M et al. Progression of human breast cancers to the metastatic state is linked berg false discovery rate procedure was performed for to genotypes of catechol-O-methyltransferase. Cancer Lett 2000; 150: multiple testing corrections.46 23–31.

The Pharmacogenomics Journal Estrogen-metabolizing gene polymorphisms and lipid levels S Almeida et al 351

12 Quinones L, Lucas D, Godoy J, Caceres D, Berthou F, Varela N et al. glutathione S-transferase (GSTM1 and GSTT1) and cytochrome P-450 CYP1A1, CYP2E1 and GSTM1 genetic polymorphisms. The effect of (CYP1A1) loci. Br J Cancer 1997; 75: 1385–1388. single and combined genotypes on lung cancer susceptibility in Chilean 30 Wang XL, Greco M, Sim AS, Duarte N, Wang J, Wilcken DE. Effect of people. Cancer Lett 2001; 174: 35–44. CYP1A1 MspI polymorphism on cigarette smoking related coronary 13 Spurr NK, Gough AC, Stevenson K, Wolf CR. Msp-1 polymorphism artery disease and diabetes. Atherosclerosis 2002; 162: 391–397. detected with a cDNA probe for the P-450 I family on chromosome 15. 31 Kvitko K, Nunes JCB, Hutz MH. (TTTA)n polymorphism of CYP19 Nucleic Acids Res 1987; 15: 5901. (aromatase gene) in Euro- and Afro-Brazilians. Genet Mol Biol 2004; 27: 14 http://www.imm.ki.se/CyPalleles. 335–336. 15 Hayashi S, Watanabe J, Nakachi K, Kawajiri K. PCR detection, of an A/G 32 Almeida S, Franken N, Zandona´ MR, Oso´rio-Wender MC, Hutz MH. polymorphism within exon 7 of the CYP1A1 gene. Nucleic Acids Res Estrogen receptor 2 and progesterone receptor gene polymorphisms 1991; 19: 4797. and lipid levels in women with different hormonal status. Pharmaco- 16 Garte S. The role of ethnicity in cancer susceptibility gene polymorph- genomics J 2005; 5: 30–34. isms: the example of CYP1A1. Carcinogenesis 1998; 19: 1329–1332. 33 Herrington DM, Howard TD, Hawkins GA, Reboussin DM, Xu J, Zheng 17 Shimada T, Yamazaki H, Mimura M, Inui Y, Guengerich FP. Inter- SL et al. Estrogen-receptor polymorphisms and effects of estrogen individual variations in human liver cytochrome P-450 enzymes replacement on high-density lipoprotein cholesterol in women with involved in the oxidation of drugs, carcinogens and toxic chemicals: coronary disease. N Engl J Med 2002; 346: 967–974. studies with liver microsomes of 30 Japanese and 30 Caucasians. 34 Koivu TA, Fan YM, Mattila KM, Dastidar P, Jokela H, Nikkari ST et al. The J Pharmacol Exp Ther 1994; 270: 414–423. effect of hormone replacement therapy on atherosclerotic severity in 18 Rebbeck TR, Jaffe JM, Walker AH, Wein AJ, Malkowicz SB. Modification relation to ESR1 genotype in postmenopausal women. Maturitas 2003; of clinical presentation of prostate tumors by a novel genetic variant in 44: 29–38. CYP3A4. J Natl Cancer Inst 1998; 90: 1225–1229. 35 Zhu BT, Conney AH. Functional role of estrogen metabolism in target 19 Hesselink DA, van Schaik RH, van der Heiden IP, van der Werf M, Gregoor cells: review and perspectives. Carcinogenesis 1998; 19:1–27. PJ, Lindemans J et al. Genetic polymorphisms of the CYP3A4, CYP3A5, 36 Bennink HJTC. Are all estrogen the same? Maturitas 2004; 47: 269–275. and MDR-1 genes and pharmacokinetics of the calcineurin inhibitors 37 Hong CC, Tang BK, Hammond GL, Tritchler D, Yaffe M, Boyd NF. cyclosporine and tacrolimus. Clin Pharmacol Ther 2003; 74: 245–254. Cytochrome P450 1A2 (CYP1A2) activity and risk factors for 20 Saito S, Iida A, Sekine A, Ogawa C, Kawauchi S, Higuchi S et al. 906 breast cancer: a cross-sectional study. Breast Cancer Res 2004; 6: variations among 27 genes encoding cytochrome P450 (CYP) enzymes R352–R365. and aldehyde dehydrogenases (ALDHs) in the Japanese population. 38 Pernerger TV. What’s wrong with Bonferroni adjustments. BMJ 1998; J Hum Genet 2002; 47: 419–444. 316: 1236–1238. 21 Tenhunen J, Salminen M, Lundstrom K, Kiviluoto T, Savolainen R, 39 American Diabetes Association. Report of the expert committee on the Ulmanen I. Genomic organization of the human catechol O-methyl- diagnosis and classification of diabetes mellitus. Diabetes Care 1997; 17: transferase gene and its expression from two distinct promoters. Eur J 1183–1201. Biochem 1994; 223: 1049–1059. 40 Friedwald WT, Levy RI, Fredrickson DS. Estimation of the concentration 22 Grossman MH, Emanuel BS, Budarf ML. Chromosomal mapping of of low-density lipoprotein cholesterol in plasma, without use of the the human catechol-O-methyltransferase gene to 22q11.1–q11.2. preparative ultracentrifuge. Clin Chem 1972; 18: 499–502. Genomics 1992; 12: 822–825. 41 Lahiri DK, Nurnberger Jr JI. A rapid non-enzymatic method for the 23 Lotta T, Vidgren J, Tilgmann C, Ulmanen I, Melen K, Julkunen I et al. preparation of HMW DNA from blood for RFLP studies. Nucleic Acids Res Kinetics of human soluble and membrane-bound catechol O-methyl- 1991; 19: 5444. transferase: a revised mechanism and description of the thermolabile 42 Kunugi H, Nanko S, Ueki A, Otsuka E, Hattori M, Hoda F et al. High variant of the enzyme. Biochemistry 1995; 34: 4202–4210. and low activity alleles of catechol-O-methyltransferase gene: ethnic 24 Berstein LM, Imyanitov EN, Kovalevskij AJ, Maximov SJ, Vasilyev DA, difference and possible association with Parkinson’s disease. Neurosci Buslov KG et al. CYP17 and CYP19 genetic polymorphisms in Lett 1997; 221: 202–204. endometrial cancer: association with intratumoral aromatase activity. 43 Tsuchiya Y, Sato T, Kiyohara C, Yoshida K, Ogoshi K, Nakamura K et al. Cancer Lett 2004; 207: 191–196. Genetic polymorphisms of cytochrome P450 1A1 and risk of gallbladder 25 http://snp500cancer.nci.nih.gov/snp.cfm?both_snp_id ¼ CYP1A2–11. cancer. J Exp Clin Cancer Res 2002; 21: 119–124. 26 Palmatier MA, Kang AM, Kidd KK. Global variation in the frequencies of 44 van Schaik RH, de Wildt SN, van Iperen NM, Uitterlinden AG, van den functionally different catechol-O-methyltransferase alleles. Biol Psychia- Anker JN, Lindemans J. CYP3A4-V polymorphism detection by try 1999; 46: 557–567. PCR-restriction fragment length polymorphism analysis and its 27 Burim RV, Canalle R, Martinelli AL, Takahashi CS. Polymorphisms in allelic frequency among 199 Dutch Caucasians. Clin Chem 2000; 46: glutathione S-transferases GSTM1, GSTT1 and GSTP1 and cytochromes 1834–1836. P450 CYP2E1 and CYP1A1 and susceptibility to cirrhosis or pancreatitis 45 Arvanitis DA, Koumantakis GE, Goumenou AG, Matalliotakis IM, in alcoholics. Mutagenesis 2004; 19: 291–298. Koumantakis EE, Spandidos DA. CYP1A1, CYP19,andGSTM1 poly- 28 Gaspar PA, Kvitko K, Papadopolis LG, Hutz MH, Weimer TA. High morphisms increase the risk of endometriosis. Fertil Steril 2003; 79: frequency of CYP1A1*2C allele in Brazilian populations. Hum Biol 2002; 702–709. 74: 235–242. 46 Benjamini Y, Hochberg Y. Controlling the false discovery rate: a 29 Esteller M, Garcia A, Martinez-Palones JM, Xercavins J, Reventos J. practical and powerful approach to multiple testing. JR Stat Soc 1995; Susceptibility to endometrial cancer: influence of allelism at p53, 57: 289–300.