Maturitas 54 (2006) 119–126

ESR1 and APOE gene polymorphisms, serum lipids,and hormonal replacement therapy

Silvana Almeida a, Marilu Fiegenbaum a, Fabiana M. de Andrade a,Maria C. Osorio-Wender b, Mara H. Hutz a,∗

a Departamento de Genetica, Instituto de Biociencias, Universidade Federal do Rio Grande do Sul,Caixa Postal 15053, 91501-970 Porto Alegre, RS, Brazil

b Departamento de Ginecologia e Obstetrıcia, Hospital de Clınicas de Porto Alegre,Universidade Federal do Rio Grande do Sul, Porto Alegre, RS, Brazil

Received 5 April 2005; received in revised form 15 September 2005; accepted 20 September 2005

Abstract

Objectives: The risks and benefits of hormone replacement therapy (HRT) are, at least in part, mediated by the metabolicindividuality of women. Therefore, we investigated the association between polymorphisms at the estrogen receptor 1 gene(ESR1) and at the apolipoprotein E gene (APOE) with lipid and lipoprotein levels in order to verify whether these concentrationsare modulated by these gene variants in women with different hormonal status.Methods: One hundred and eighteen postmenopausal women using oral HRT with estrogen or estrogen plus progestagen (HRT+,mean age = 56 ± 6.7 years, 39–75 years) and 167 postmenopausal women that were not on HRT (HRT−, mean age = 58 ± 9.8

years, 38–85 years) participated in the study. The polymorphisms were genotyped by PCR-RFLP methods.Results: No significant effect of ESR1 genotypes or haplotypes and ESR1*HRT interactions were detected on lipid levels intwo-way analysis of variance. Postmenopausal women HRT nonusers carriers of the APOE*4 allele had higher T-chol andLDL-C levels than postmenopausal women HRT nonusers carriers of the APOE*3 and APOE*2 allele. T-chol and LDL-Cconcentrations in postmenopausal users of HRT that were APOE*4 carriers were similar to those in postmenopausal womennonusers of HRT homozygotes for APOE*3 and APOE*2 carriers. A significant APOE*4/HRT interaction was detected onT-chol and LDL-C levels by multiple regression analysis.Conclusion: The results from this study suggest that the HRT influence on T-chol and LDL-C levels is modulated by APOEisoforms but not by ESR1 polymorphisms.© 2005 Elsevier Ireland Ltd. All rights reserved.

Keywords: Estrogen; Estrogen receptor 1 gene; Apolipoprotein E gene; Lipid levels; Hormonal replacement therapy

∗ Corresponding author. Tel.: +55 51 3316 6720;fax: +55 51 3343 5850.

E-mail address: mara.hutz@ufrgs.br (M.H. Hutz).

1. Introduction

The role of estrogen replacement therapy (ERT)and/or hormone replacement therapy (HRT) in primary

0378-5122/$ – see front matter © 2005 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.maturitas.2005.09.009

120 S. Almeida et al. / Maturitas 54 (2006) 119–126

and secondary prevention of cardiovascular disease(CVD) remains controversial [1]. The discrepan-cies between observational and randomized datahave prompted an intense discussion, and severalphysiological issues have also been explored. Theeffects of HRT on lipids, haemostatic parametersand vascular wall support the beneficial cardiovas-cular effect. Treatment with estrogen lowers plasmalevels of total cholesterol (T-chol) and low-densitylipoprotein cholesterol (LDL-C), and increaseshigh-density lipoprotein cholesterol (HDL-C) [2].In addition, estrogen accelerates endothelial cellgrowth, inhibits the migration and proliferation ofsmooth-muscle cells in vitro, increases vasodilatation,inhibits the response of blood vessels to injury anddevelopment of atherosclerosis [3]. Nevertheless,some studies have demonstrated a rise in C-reactiveprotein (CRP) levels after initiation of oral HRT[4].

One interesting question concerning the cardiovas-cular effects of HRT is whether genetic factors canmodify, at least in part, individual responses to estro-gen treatment. Knowledge of such genetic factors mayalso contribute to the biological variability in manyother physiological processes influenced by exogenousestrogen; however, only few studies reported the impactof genetic factors on lipid responses to HRT/ERT[5–16].

An important polymorphism associated withcholesterol levels is apolipoprotein E (APOE). Ithwvvwglopfpammg[cg

al. [23] demonstrated that the association of APOEgenotypes with lipoproteins is significantly greaterin postmenopausal than in premenopausal women incross-sectional and follow-up studies, respectively.Several investigations have studied the relationshipbetween APOE genotypes and lipid levels in the gen-eral population, but only few studies have addressedthe relation between HRT and APOE genotypes[5–9,11–12]. Although some of these studies indicatethat APOE genotypes contribute to differential varia-tion in lipid levels with menopause and HRT [6–9,11],this variation was not completely explained by APOEgenotypes. Since estrogen receptors are expressed inmost cardiovascular tissues [24–26] and appear toplay an important role in the expression of Apo AI, theprimary protein constituent of HDL-C [27], and in theseverity of atherosclerotic lesions [28], polymorphismsin estrogen receptor genes (ESR) are also naturalcandidates for association with lipid levels and HRTeffect.

In this cross-sectional study, we investigated theassociation between four common single nucleotidepolymorphisms (SNPs) in the estrogen receptor 1(ESR1) gene locus, −1989T>G, +261G>C, IVS1-397T>C and IVS1-351A>G, and APOE isoforms withplasma lipid levels in postmenopausal women exposedand unexposed to HRT in order to verify whether theselipid levels are influenced differently by these polymor-phisms according to hormonal status.

2

2

wiGtgrpgp2fa

as been firmly established that the APOE isoforms,hich are mainly expressed on chylomicrons andery low-density lipoproteins, explains 4–15% of theariation on LDL-C concentrations [17]. Comparedith the APOE E*3 homozygotes, the most commonenotype, the E*2 allele is associated with lowerevels of cholesterol, whereas the E*4 allele haspposite effects [18]. In men the APOE*2/*3/*4olymorphism is a significant genetic risk factoror coronary atherosclerosis [19]. In addition to thisolymorphism, recently Viiri et al. [20] reported thatpromoter polymorphism as well as their haplotypesodulate lipid and lipoprotein concentrations iniddle aged Finnish men. Estrogen regulates APOE

ene expression by posttranscriptional mechanisms21], therefore it is expected that differential estrogenoncentrations might modulate the effect of APOEene polymorphisms. Schaefer et al. [22] and Hak et

. Subjects and methods

.1. Subjects

The sample consisted of 285 postmenopausalomen of European descent ascertained at two clin-

cal centers from the Universidade Federal do Riorande do Sul as previously described [29]. Briefly,

he postmenopausal women were divided in two sub-roups according to hormonal status: 118 women wereeceiving oral HRT with estrogen or estrogen plusrogestagen (0.625 mg of conjugated equine estro-en, n = 24; 0.625 mg of conjugated equine estrogenlus 2.5 mg of medroxyprogesterone acetate, n = 62;mg of estradiol plus noretisterone acetate, n = 32)

or at least four months (HRT users or HRT+, meange = 56 ± 6.7 years, 39–75 years) and 167 women

S. Almeida et al. / Maturitas 54 (2006) 119–126 121

Table 1Characteristics of the sample

HRT+ HRT−Number 118 167Age (years)a 56 ± 6.7 58 ± 9.8Time after menopause (years)a 9 ± 6.4 7 ± 6.4T-chol (mmol/l)a 5.60 ± 0.90 5.96 ± 1.18HDL-C (mmol/l)a 1.57 ± 0.36 1.24 ± 0.31LDL-C (mmol/l)a 3.38 ± 0.79 3.97 ± 1.00TG (mmol/l)a 1.40 ± 0.59 1.61 ± 0.80BMI (kg/m2)a 26.9 ± 4.4 27.7 ± 4.6Physical activity (%) 46 36Current smokers (%) 15 19Hypertension (%) 42 37

Unadjusted data. T-chol indicates total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoproteincholesterol; TG, triglycerides; BMI, body mass index; HRT, hor-monal replacement therapy; HRT+, postmenopausal women HRTusers; HRT−, postmenopausal women HRT nonusers.

a Values are mean ± S.D.

that were not using HRT (HRT nonusers or HRT−,mean age = 58 ± 9.8 years, 38–85 years) for at least 1year. The women were considered postmenopausal ifthey had no vaginal bleeding for at least 12 monthswithout other obvious pathological or physiologicalcause or bilateral oophorectomy. Information abouthealth and lifestyle factors (physical activity, smok-ing status, alcohol consumption, and drugs intake)was obtained from each individual by an interview.After answering the interview, individuals wearinglight clothes had their body weight and height recordedand their body mass index (BMI) calculated by theratio of weight (in kg) and height squared (in meters).Exclusion criteria were secondary hyperlipidemia dueto renal, hepatic or thyroid disease, and diabetes orfasting blood glucose levels above 6.9 mmol/l [30].Individuals that were on lipid-lowering medicationwere excluded. The characteristics of the individualsincluded in the sample are shown in Table 1. Informedconsent was obtained from each subject included in thesample.

2.2. Measurements of serum lipid levels

Blood samples were collected from patients after12 h fasting. T-chol, HDL-C, TG, and glucose lev-els were determined by standard methods using com-mercial kits. LDL-C was calculated according to theFriedewald et al. formula [31].

2.3. DNA analyses

Genomic DNA was extracted from peripheral bloodleukocytes by a salting out procedure [32]. DNA wasamplified by polymerase chain reaction, using thesame conditions and oligonucleotide primers as pre-viously described [33–37]. The amplification prod-ucts were subsequently digested with the followingrestriction enzymes under conditions recommendedby the manufacturer: HhaI (ESR1 −1989T>G andAPOE), BstUI (ESR1 +261G>C), PvuII (ESR1 IVS1-397T>C), and XbaI (ESR1 IVS1-351A>G). Genotypeswere determined after electrophoresis on agarose gelscontaining ethidium bromide, using a 100 bp ladder toscore the band sizes, and to identify the correspondingalleles.

2.4. Statistical analyses

Allele frequencies were estimated by gene count-ing. A χ2-test for goodness of fit was used to verifywhether the observed allele frequencies agreed withthose expected under Hardy–Weinberg equilibrium.The maximum likelihood estimate of haplotype fre-quencies were calculated from multisite marker datausing the Multiple Locus Haplotype Analysis ver. 2.0[38–40]. Linkage disequilibrium was tested by a χ2

goodness of fit test using Arlequin program version2.000 [41].

All other tests and transformations were performedwwmvaatmTslpEEEEnt

ith SPSS 8.0 statistical package. TG levels and BMIere ln-transformed to remove skewness, geometriceans are shown in tables. The influences of ESR1

ariation on lipids as well as the genotype by HRT inter-ction on lipid levels were performed by a two-waynalysis of variance. Student’s t-test was performedo access differences between HRT+ and HRT− post-

enopausal women for T-chol, HDL-C, LDL-C andG concentrations within each APOE group. Analy-es of variance (ANOVA) were employed to compareipid levels between APOE groups within each sam-le. In these analyses, patients carriers of E*2/E*2 and*2/E*3 genotypes were denominated E2, while the*3/E*3 genotype was designated as the E3 group, and*3/E*4 and E*4/E*4 genotypes were joined in the4 group; the patients with E*2/E*4 genotype wereot included in the analyses because we choose noto assign these patients to either group E2 or E4. The

122 S. Almeida et al. / Maturitas 54 (2006) 119–126

interaction between the APOE alleles and HRT on lipidlevels was tested by multiple regression analyses, ageand BMI entered the model as covariates, the refer-ence group was non HRT users carriers of the APOEE*3/E*3 genotype.

3. Results

3.1. Allele, genotype, and haplotype frequencies

In the study population, allele frequencies ofthe ESR1 gene polymorphisms were −1989T>G*T(0.84), +261G>C*G (0.95), IVS1-397T>C*T (0.57),and IVS1-351A>G*A (0.66). The allele frequen-cies of APOE gene polymorphism were E*2 (0.06),E*3 (0.81), and E*4 (0.13). The genotype fre-quencies observed did not show statistically signif-icant differences compared to those expected underHardy–Weinberg equilibrium. Thirteen haplotypes ofthe ESR1 gene polymorphisms were detected inthe sample, the two most common haplotypes were−1989T>G*T – +261G>C*G – IVS1-397T>C*T –IVS1-351A>G*A and −1989T>G*T – +261G>C*G –IVS1-397T>C*C – IVS1-351A>G*G which accountedfor 47 and 31% of the investigated chromosomes,respectively.

3.2. Association of ESR1 and APOE genotypeswith lipid and lipoprotein levels

No isolated effect of the four ESR1 gene polymor-phisms (data not shown), or haplotypes as well theinteraction between these genetic variants with HRTuse was detected (Table 2).

Table 3 presents lipid and lipoprotein levels byAPOE genotypes in HRT users (HRT+) and nonusers(HRT−). An association between lipid levels andAPOE isoforms were observed only in HRT nonusers.In this group, E*4 carriers showed higher total choles-terol concentrations than E*3 homozygotes (TukeyHSD p = 0.006). LDL-C levels were also higher inE*4 carriers than those observed in E*2 (Tukey HSDp = 0.010) and E*3 (Tukey HSD p = 0.002) groups.

Fig. 1 compares the HRT effect by genotype.Women in E2 and E3 groups have similar total choles-terol levels independently of HRT usage, whereasAPOE*4 carriers that were on HRT use showed 10%less T-chol than nonusers (p = 0.049; Fig. 1A). LDL-Cconcentrations were also lower in HRT users homozy-gous for E3 and E4 carriers when compared to theirHRT nonuser counterparts (−13%, p < 0.0001 and−21%, p = 0.0008, respectively — Fig. 1B). A signifi-cant interaction between APOE*4 and HRT on T-chollevels (p < 0.001; �-coefficient = 0.720); and LDL-C

Table 2T oproteiI *A) anda

HR

M

TH 5.9A 5.9

LH 3.9A 3.9

HH 1.2A 1.2

TH 1.5A 1.6

T holesterH n HRT

otal cholesterol, low-density lipoprotein cholesterol, high-density lip(−1989T>G*T – +261G>C*G – IVS1-397T>C*T – IVS1-351A>Gnd unexposed to hormonal replacement therapy

HRT+

Mean ± S.D. n

-chol (mmol/l)omozygotes I 5.52 ± 0.86 33ll others 5.63 ± 0.91 86

DL-C (mmol/l)omozygotes I 3.37 ± 0.67 33ll others 3.37 ± 0.83 86

DL-C (mmol/l)omozygotes I 1.50 ± 0.37 33ll others 1.61 ± 0.36 86

G (mmol/l)omozygotes I 1.40 ± 0.47ll others 1.40 ± 0.63

-chol indicates, total cholesterol; LDL-C, low-density lipoprotein cRT, hormonal replacement therapy; HRT+, postmenopausal wome

n cholesterol, and triglycerides means in homozygotes for haplotypecarriers of all others haplotypes in postmenopausal women exposed

T- p

ean ± S.D. n

HRT = 0.0106 ± 1.14 40 Genotype = 0.7366 ± 1.20 127 HRT* genotype = 0.691

HRT < 0.0019 ± 0.97 40 Genotype = 0.9527 ± 1.02 124 HRT* genotype = 0.950

HRT < 0.0017 ± 0.27 40 genotype = 0.5643 ± 0.33 126 HRT* genotype = 0.129

HRT = 0.2303 ± 0.69 40 Genotype = 0.6444 ± 0.84 125 HRT* genotype = 0.522

ol; HDL-C, high-density lipoprotein cholesterol; TG, triglycerides;users; HRT−, postmenopausal women HRT nonusers.

S. Almeida et al. / Maturitas 54 (2006) 119–126 123

Table 3Total cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and triglycerides means for postmenopausal womenreceiving and not receiving hormonal replacement therapy according to the APOE genotype

HRT+ HRT−E2 (n = 12) E3 (n = 85) E4 (n = 20) p E2 (n = 18) E3 (n = 106) E4 (n = 39) p

T-chol (mmol/l) 5.47 ± 1.30 5.55 ± 0.84 5.85 ± 0.78 0.348 5.83 ± 1.13b 5.78 ± 1.12b 6.47 ± 1.26b 0.007LDL-C (mmol/l) 3.04 ± 1.07 3.35 ± 0.74 3.56 ± 0.72 0.188 3.66 ± 0.96c 3.87 ± 0.94c 4.49 ± 1.05c 0.001HDL-C (mmol/l) 1.62 ± 0.33 1.60 ± 0.36 1.57 ± 0.36 0.925 1.29 ± 0.44 1.24 ± 0.29 1.21 ± 0.29 0.755TG (mmol/l)a 1.50 ± 0.02 1.22 ± 0.01 1.45 ± 0.02 0.051 1.33 ± 0.02 1.41 ± 0.02 1.53 ± 0.02 0.477

T-chol indicates, total cholesterol; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; TG, triglycerides;BMI, body mass index. E2 included individuals with E*2/E*2 and E*2/E*3 genotypes; group E3, E*3/E*3; and group E4, E*3/E*4 and E*4/E*4genotypes.

a Geometric mean.b Multiple comparisons Tukey HSD test, E4 and E2 p = 0.119; E4 and E3 p = 0.006.c Multiple comparisons Tukey HSD test, E4 and E2 p = 0.010; E4 and E3 p = 0.002.

levels (p < 0.001; �-coefficient = 0.650) was detectedby regression analyses after adjustment for age andBMI.

4. Discussion

The observed allele frequencies for +261G>C,IVS1-397T>C, and IVS1-351A>G polymorphismsof ESR1 and APOE polymorphism in the sam-ple were similar to those reported in other Euro-pean or European-derived populations [10,18,33,34,36,42–44]. The ESR1 −1989T>G polymorphism wasinvestigated in Japanese patients only [37].

Only few previous association studies investigatedESR1 gene polymorphisms with lipoprotein levels.Matsubara et al. [45] did not detect an associationbetween IVS1-397T>C and IVS1-351A>G polymor-phisms and lipoprotein fractions in Japanese men andpostmenopausal women. Herrington et al. [10] reportedan association between IVS1-397T>C and response ofHDL-C level to hormone replacement therapy in post-menopausal women. The T/T genotype of the IVS1-397T>C polymorphism was associated with higher lev-els of T-chol and LDL-C in an adolescent female cohort[14]. It has also been reported that the ESR1 gene is anindependent predictor of severe atherosclerotic lesionsin Finnish males and females [16,28]. These studiessuggested that the role of ESR1 gene in lipoproteinmetabolism remain uncertain. The overall conclusionoEm

therefore detected only in follow-up studies or otherconditions in which environment factors are more con-trolled.

Although several works have studied the relation-ship between APOE genotypes and lipid levels, andatherosclerosis in the general population [19,20],only few investigations have addressed hormonalreplacement therapy and APOE associations with thelipid profile of postmenopausal women. In follow-upstudies, three investigations described an effect ofAPOE genotypes in response to hormonal replacementtherapy [7–9], while two other HRT follow-up studiesdid not report significant differences between APOEgenotypes and response to lipid and lipoprotein levels[5,12]. Von Muhlen et al. [11] described E2 genotypesassociations with lower T-chol and LDL-C levels anda significant association between the APOE–E2*ERTinteraction and HDL-C, while Garry et al. [6] demon-strated that E4 women on ERT had significantly lowerT-chol than women with the same genotype thatwere not on ERT. Similar results were observed in thepresent study (Fig. 1A and B). The present results takentogether with those previously described by Garry etal. [6] suggest that hormone therapy may have a rolein determining APOE allele effects. These effects alsosuggest that women carriers of E*4 allele may benefitmore from HRT than women with other APOE geno-types. These results taken together with those of Fan etal. [15] for a hepatic lipase gene polymorphism help usunderstand the genetic effects on the risk and benefitso

t

f this study is of an absence of association between theSR1 gene and lipid levels, or the effect of these poly-orphisms on HRT response may be very small and

f HRT.In conclusion, the results from this study suggest

hat the HRT influence on T-chol and LDL-C levels is

124 S. Almeida et al. / Maturitas 54 (2006) 119–126

Fig. 1. Summary of total cholesterol — T-chol (A) and low-densitylipoprotein cholesterol LDL-C (C) levels in postmenopausal womenreceiving hormonal replacement therapy (HRT+: white bars) or notreceiving (HRT−: gray bars). (A) The within APOE group dif-ferences between HRT− and HRT+ women for T-chol were notsignificant for E2 (−6%, p = 0.425) and E3 groups (−4%, p = 0.115),but was significant for E4 group (−10%, p = 0.049). (B) The withinAPOE group differences between HRT− and HRT+ for LDL-C wasnot significant for E2 group (−17%, p = 0.112), but were significantfor groups E3 (−13%, p = 0.0001) and E4 (−21%, p = 0.0008).

modulated by APOE genotypes but not by ESR1 poly-morphisms. This study is limited by its cross-sectionaldesign and clearly, further association and prospectivefollow-up studies in larger samples are warrantedbefore definitive conclusions are reached about therole of these polymorphisms on lipid fractions inwomen on HRT.

Acknowledgements

Thanks are due to Ana Lucia S. Antunes, and MariaPerpetua de O. Pinto from the Clinical Analysis Lab-

oratory of the Pharmacy College. We are also gratefulto Fabiano Roldao Silveira and Marcel Arsand for helpin sample collection, and to Dr. Sidia M. Callegari-Jacques for her advice on the statistical analysis. Finan-cial support was provided by Programa de Apoio aNucleos de Excelencia (PRONEX, Brazil) and Con-selho Nacional de Desenvolvimento Cientıfico e Tec-nologico (CNPq, Brazil).

References

[1] Benagiano G, Farris M. Why a consensus conference on hor-mone replacement therapy and the cardiovascular system?Maturitas 2004;47:245–53.

[2] Godsland IF. Effects of postmenopausal hormone replacementtherapy on lipid, lipoprotein, and apolipoprotein (a) concentra-tions: analysis of studies published from 1974 to 2000. FertilSteril 2001;75:898–915.

[3] Mendelsohn ME, Karas RH. The protective effects of estrogenon the cardiovascular system. N Engl J Med 1999;340:1801–11.

[4] Maas AH, van der Schouw YT, Grobbee DE, van der GraafY. Rise and fall of hormone therapy in postmenopausalwomen with cardiovascular disease. Menopause 2004;11:228–35.

[5] Somekawa Y, Wakabayashi A. Relationship betweenapolipoprotein E polymorphism, menopausal symptoms, andserum lipids during hormone replacement therapy. Eur J ObstetGynecol Reprod Biol 1998;79:185–91.

[6] Garry PJ, Baumgartner RN, Brodie SG, et al. Estrogenreplacement therapy, serum lipids, and polymorphism of theapolipoprotein E gene. Clin Chem 1999;45:1214–23.

[7] Heikkinen AM, Niskanen L, Ryynanen M, et al. Is the response

[

[

[

of serum lipids and lipoproteins to postmenopausal hormonereplacement therapy modified by ApoE genotype? ArteriosclerThromb Vasc Biol 1999;19:402–7.

[8] Tolosa M, Peiro S, Real JT, Cano A, Ascaso JF, CarmenaR. The influence of apo E phenotypes on the plasma triglyc-erides response to hormonal replacement therapy during themenopause. Maturitas 2001;40:173–81.

[9] Tsuda M, Sanada M, Nakagawa H, Kodama I, Sakashita T,Ohama K. Phenotype of apolipoprotein E influences the lipidmetabolic response of postmenopausal women to hormonereplacement therapy. Maturitas 2001;38:297–304.

10] Herrington DM, Howard TD, Hawkins GA, et al. Estrogen-receptor polymorphisms and effects of estrogen replacementon high-density lipoprotein cholesterol in women with coronarydisease. N Engl J Med 2002;346:967–74.

11] von Muhlen D, Barrett-Connor E, Kritz-Silverstein D.Apolipoprotein E genotype and response of lipid levels to post-menopausal estrogen use. Atherosclerosis 2002;161:209–14.

12] Lehtimaki T, Dastidar P, Jokela H, et al. Effect of long-termhormone replacement therapy on atherosclerosis progressionin postmenopausal women relates to functional apolipoproteinE genotype. J Clin Endocrinol Metab 2002;87:4147–53.

S. Almeida et al. / Maturitas 54 (2006) 119–126 125

[13] Yamakawa-Kobayashi K, Somekawa Y, Fujimura M, TomuraS, Arinami T, Hamaguchi H. Relation of the −514C/T poly-morphism in the hepatic lipase gene to serum HDL and LDLcholesterol levels in postmenopausal women under hormonereplacement therapy. Atherosclerosis 2002;162:17–21.

[14] Nordstrom P, Glader CA, Dahlen G, et al. Oestrogen receptoralpha gene polymorphism is related to aortic valve sclerosis inpostmenopausal women. J Intern Med 2003;254:140–6.

[15] Fan Y-M, Dastidar P, Jokela H, Punnonen R, Lehtimaki. T.review: Hepatic lipase C-480T genotype-dependent benefitfrom long-term hormone replacement therapy for atheroscle-rosis progression in postmenopausal women. J Clin EndocrinolMetab 2005;90:3786–92.

[16] Koivu TA, Fan Y-M, Mattila KM, et al. The effect of hor-mone replacement therapy on atherosclerotic severity in rela-tion to ESR1 genotype in postmenopausal women. Maturitas2003;44:29–38.

[17] Kamboh MI, Aston CE, Ferrell RE, Hamman RF. Impact ofapolipoprotein E polymorphism in determining interindivid-ual variation in total cholesterol and low-density lipoproteincholesterol in Hispanics and non-Hispanic whites. Atheroscle-rosis 1993;98:201–11.

[18] Mahley RW, Rall Jr SC. Apolipoprotein E: far more thana lipid transport protein. Annu Rev Genomics Hum Genet2000;1:507–37.

[19] Ilveskoski E, Perola M, Lehtimaki T, et al. Age-dependent asso-ciation of apolipoprotein E genotype with coronary and aorticatherosclerosis in middle-aged men. An autopsy study. Circu-lation 1999;100:608–13.

[20] Viiri LE, Loimaala A, Nenonen A, et al. The association ofapolipoprotein E gene promoter polymorphisms and haplotypeswith serum lipid and lipoprotein concentrations. Atherosclero-sis 2005;179:161–7.

[21] Srivastava RA, Srivastava N, Averna M, et al. Estrogen up-regulates apolipoprotein E (APOE) gene expression by increas-

[

[

[

[

[

[27] Lamon-Fava S. Genistein activates apolipoprotein A-I geneexpression in the human hepatoma cell line Hep G2. J Nutr2000;130:2489–92.

[28] Lehtimaki T, Kunnas TA, Mattila KM, et al. Coronary arterywall atherosclerosis in relation to the estrogen receptor1gene polymorphism: an autopsy study. J Mol Med 2002;80:176–80.

[29] Almeida S, Franken N, Zandona MR, Osorio-Wender MC, HutzMH. Estrogen receptor 2 and progesterone receptor gene poly-morphisms and lipid levels in women with different hormonalstatus. Pharmacogenomics J 2005;5:30–4.

[30] American Diabetes Association. Report of the expert committeeon the diagnosis and classification of diabetes mellitus. DiabetesCare 1997;20:1183–1201.

[31] Friedwald WT, Levy RI, Fredrickson DS. Estimation of theconcentration of low-density lipoprotein cholesterol in plasma,without use of the preparative ultracentrifuge. Clin Chem1972;18:499–502.

[32] Lahiri DK, Nurnberger Jr JI. A rapid non-enzymatic method forthe preparation of HMW DNA from blood for RFLP studies.Nucleic Acids Res 1991;19:5444.

[33] Andersen TI, Heimdal KR, Skrede M, Tveit K, Berg K, Borre-sen AL. Oestrogen receptor (ESR) polymorphisms and breastcancer susceptibility. Hum Genet 1994;94:665–70.

[34] Kobayashi S, Inoue S, Hosoi T, Ouchi Y, Shiraki M, OrimoH. Association of bone mineral density with polymorphismof the estrogen receptor gene. J Bone Miner Res 1996;11:306–11.

[35] Yaich L, Dupont WD, Cavener DR, Parl FF. Analysis ofthe PvuII restriction fragment-length polymorphism and exonstructure of the estrogen receptor gene in breast cancer andperipheral blood. Cancer Res 1992;52:77–83.

[36] de-Andrade FM, Larrandaburu M, Callegari-Jacques SM,Gastaldo G, Hutz MH. Association of apolipoprotein Epolymorphism with plasma lipids and Alzheimer’s disease

[

[

[

[

[

[

ing APOE mRNA in the translating pool via the estrogen recep-tor alpha-mediated pathway. J Biol Chem 1997;272:33360–6.

22] Schaefer EJ, Lamon-Fava S, Johnson S, et al. Effects of gen-der and menopausal status on the association of apolipopro-tein E phenotype with plasma lipoprotein levels. Resultsfrom the Framingham Offspring Study. Arterioscler Thromb1994;14:1105–13.

23] Hak AE, Witteman JC, Hugens W, et al. The increase in choles-terol with menopause is associated with the apolipoprotein Egenotype. A population-based longitudinal study. Atheroscle-rosis 2004;175:169–76.

24] Karas RH, Patterson BL, Mendelsohn ME. Human vascularsmooth muscle cells contain functional estrogen receptor. Cir-culation 1994;89:1943–50.

25] Losordo DW, Kearney M, Kim EA, Jekanowski J, Isner JM.Variable expression of the estrogen receptor in normal andatherosclerotic coronary arteries of premenopausal women. Cir-culation 1994;89:1501–10.

26] Venkov CD, Rankin AB, Vaughan DE. Identification of authen-tic estrogen receptor in cultured endothelial cells. A potentialmechanism for steroid hormone regulation of endothelial func-tion. Circulation 1996;94:727–33.

in a Southern Brazilian population. Braz J Med Biol Res2000;33:529–37.

37] Lu H, Higashikata T, Inazu A, et al. Association of estrogenreceptor-alpha gene polymorphisms with coronary artery dis-ease in patients with familial hypercholesterolemia. ArteriosclerThromb Vasc Biol 2002;22:817–23.

38] Long JC, Williams RC, Urbanek M. An E–M algorithm andtesting strategy for multiple-locus haplotypes. Am J Hum Genet1995;56:799–810.

39] Long JC. Multiple locus haplotype analysis, version 2.0. Soft-ware and documentation distributed by the author. Sectionon population genetics and linkage, Bethesda, MD: Labora-tory of Neurogenetics, niaaa, National Institutes of Health;1999.

40] Peterson RJ, Goldman D, Long JC. Effects of worldwide popu-lation subdivision on ALDH2 linkage disequilibrium. GenomeRes 1999;9:844–52.

41] Schneider S, Kueffer J-M, Roessli D, Excoffier L. Arlequin ver2.000: a software for population data analysis. Geneva: Geneticsand Biometry Laboratory, University of Geneva; 2000.

42] Vandevyver C, Vanhoof J, Declerck K, et al. Lack of associationbetween estrogen receptor genotypes and bone mineral density,

126 S. Almeida et al. / Maturitas 54 (2006) 119–126

fracture history, or muscle strength in elderly women. J BoneMiner Res 1999;14:1576–82.

[43] Weel AE, Uitterlinden AG, Westendorp IC, et al. Estrogenreceptor polymorphism predicts the onset of natural and sur-gical menopause. J Clin Endocrinol Metab 1999;84:3146–50.

[44] Deng HW, Li J, Li JL, et al. Association of estrogen receptor-alpha genotypes with body mass index in normal healthy

postmenopausal Caucasian women. J Clin Endocrinol Metab2000;85:2748–51.

[45] Matsubara Y, Murata M, Kawano K, et al. Genotype distri-bution of estrogen receptor polymorphisms in men and post-menopausal women from healthy and coronary populations andits relation to serum lipid levels. Arterioscler Thromb Vasc Biol1997;17:3006–12.