Association of Estrogen Receptor-α Gene Polymorphisms With Coronary Artery Disease in Patients With Familial Hypercholesterolemia
To investigate the association of estrogen receptor (ER)-α gene polymorphisms with coronary artery disease (CAD), we studied 197 men and 98 postmenopausal women with heterozygous familial hypercholesterolemia. We examined the known polymorphisms, including PvuII, XbaI, TA repeat, and CA repeat, and identified 6 novel polymorphisms in the ER-α gene. The distributions of −1989T/G (a novel polymorphism in promoter B) and XbaI in intron 1 were associated with CAD in postmenopausal women and in men, with a higher frequency of the G/G genotype (P=0.03) or X1/X1 genotype (P=0.02) in the CAD group. The frequency of alleles of TA repeats >17 was found to be significantly higher in postmenopausal women with CAD than in those without CAD (P=0.04), but not in men. Logistic regression analysis with all coronary risk factors as covariates showed that the G/G genotype was a higher risk for CAD (odds ratio 4.5, 95% CI 1.0 to 19.5;P=0.04) but that X1/X1 was not. We conclude that −1989T/G or its linked polymorphisms in the ER-α gene may confer risk for CAD and that the G/G genotype may be an independent predictor for CAD in patients with familial hypercholesterolemia.
Estrogens, the effects of which are known to be mediated primarily via binding to estrogen receptor (ER)-α and ER-β, have been reported to provide protection against coronary artery disease (CAD) in postmenopausal women.1 ER-α and ER-β have 6 domains (A through F) encoded by 8 exons. ER-α has an activation function-1 (AF-1) in the A/B domain that is lacking in ER-β, although the DNA- and ligand-binding domains are highly conserved in both ERs.2 Five polymorphisms in the ER-α gene have been reported in the genomic DNA extracted from human breast tumors or normal human peripheral blood leukocytes.3–11⇓⇓⇓⇓⇓⇓⇓⇓ The first polymorphism, PvuII, caused by a C/T transition (P1/P2) in intron 1, is located ≈0.4 kb upstream from exon 2.3 The second polymorphism, XbaI, caused by a G/A transition (X1/X2), is located ≈50 bp downstream from the PvuII polymorphic site. Recent studies have suggested that the 2 polymorphisms might be associated with breast cancer3,4⇓ and postmenopausal osteoporosis.5,6⇓ The third polymorphism B variant, caused by a G/C transversion at codon 87 in exon 1, has been reported to be associated with hypertension.7 Another 2 polymorphisms are dinucleotide (TA and CA) repeat polymorphisms. The TA repeat is located at the 5′ untranslated region, the length of which has associated with postmenopausal osteoporosis8,9⇓ and the severity of CAD in white men.10 The CA repeat, which was newly defined in intron 5,11 is the marker D6S440.12
Smith et al13 have reported a T/C transition at codon 157 of both alleles in exon 2 of the ER-α gene in a 28-year-old man, resulting in a premature stop codon. Subsequently, he had evidence of early atherosclerosis at the age of 31 years.14 This observation prompted us to test the hypothesis that genetic variations in the ER-α gene could be a risk factor for CAD. Familial hypercholesterolemia (FH) is an autosomal dominant disorder characterized by primary hypercholesterolemia and a higher incidence of premature CAD. Heterozygotes occur in the population at a frequency of ≈1 in 500, placing FH among the most common monogenic diseases in humans. Women with FH suffer clinical CAD later than affected men. In England, Slack15 found the mean age of onset of CAD to be 43 years for men and 53 years for women. In the United States, Stone et al16 found that the cumulative probability of CAD was 52% in male heterozygotes and 33% in female heterozygotes aged 60 years. Mabuchi et al17 reported that the incidence of CAD was 43%, and earlier death from CAD was found in male Japanese heterozygotes (aged 54 years) compared with female heterozygotes (aged 68 years). Interestingly, this sex difference, which is also a characteristic feature of the usual form of CAD in general populations, does not seem to be operative in homozygotes.18 It is a unique approach to investigate the role of ER-α gene polymorphisms in a group with heterozygous FH that has a major coronary risk factor and sex difference in the risk of CAD. Therefore, we investigated the association of the known and 6 novel polymorphisms in the ER-α gene with CAD in patients with heterozygous FH.
Two hundred ninety-five unrelated Japanese patients with heterozygous FH who were dwelling in the Hokuriku area and recruited at Kanazawa University Hospital were analyzed: 119 with CAD, either myocardial infarction (MI, 48 men and 9 women) or angina pectoris (47 men and 15 women), and 176 without CAD (102 men and 74 women). The study was approved by the Ethics Committee of Kanazawa University Medical School. Heterozygous FH was diagnosed when either of 2 sets of criteria was met: (1) primary hypercholesterolemia (total cholesterol ≥5.9 mmol/L and <12.9 mmol/L) with tendon xanthomas or (2) primary hypercholesterolemia with the definite diagnosis of FH in any first-degree relative.19 Diagnosis of MI was based on clinical symptoms, appropriate new onset of ECG changes, and elevated serum creatine phosphokinase levels. All but 3 patients with MI were evaluated by coronary angiography. Diagnosis of angina pectoris was made according to clinical symptoms and ECG findings, and only those patients whose coronary lesions were confirmed by coronary angiography were eligible for the present study. Those who were diagnosed with vasospastic angina pectoris, according to clinical symptoms and ECG findings with normal coronary angiogram, were excluded. To determine the severity of CAD, we used 2 systems: (1) affected vessel numbers and (2) the coronary stenosis index. In the former system, the severity of CAD was expressed simply by affected vessel numbers (≥75% stenosis of lumen size), classifying numbers between 1 and 3. The lesion at left main coronary artery was regarded as 2-vessel disease. In the latter system, the severity of stenotic changes was assessed by a score assigned to each of 15 segments as described elsewhere.19 For patients with MI, ages at their first events were recorded, whereas for patients with angina pectoris, ages at which coronary angiography was performed were recorded. Non-CAD patients were characterized by no history of angina and other heart diseases, a normal resting ECG, and normal exercise ECG stress testing. Half of non-CAD patients (89 of 176) received coronary angiographic examination. No obvious stenosis (0% to 50% stenosis of lumen size) was found. Patients who were aged >80 years or had thyroid disease, breast cancer, or uterine or ovarian tumor were excluded. No individual was receiving or had previously received hormone replacement therapy. All women were postmenopausal, as defined by the absence of menstruation for >6 months or having attained an age ≥60 years. Women with surgical menopause were excluded. Clinical data, including body mass index, smoking history, blood pressure, diabetes status, and lipid profiles, were collected before introducing lipid-lowering therapy. Hypertension was defined as being present if antihypertensive treatment has been instituted or if consecutive 2-visit measurements of blood pressure were >160 mm Hg systolic or 95 mm Hg diastolic. Manifest diabetes mellitus was defined as present with fasting plasma glucose ≥140 mg/dL or ≥200 mg/dL at 120 minutes after 75 g of oral glucose loading.19 Subjects who smoked ≥10 cigarettes/d were classified as current smokers.
Determination of ER-α Genotypes
Blood was obtained from peripheral veins of patients after informed consent. PvuII, XbaI, and B-variant polymorphisms were analyzed by polymerase chain reaction (PCR) restriction fragment length polymorphism methods. For PvuII and XbaI polymorphisms, a 1.3-kb DNA fragment of the ER-α gene that contains the 2 polymorphic sites was amplified by PCR, 3,4⇓ and for B-variant polymorphism, a 195-bp DNA fragment was amplified.4 PCR products were digested with PvuII or XbaI (Takara Shuzo) for PvuII or XbaI polymorphism or AccII (Nippon Gene) for B-variant polymorphism at 37°C overnight, to determine genotypes.
Two dinucleotide repeat polymorphisms of the ER-α gene were investigated by PCR to amplify the TA repeat, which is located ≈1.2 kb upstream from exon 1 (relative to the first transcribed nucleotide),9 and the CA repeat, which is located 224 bp downstream from exon 5.11 The primers were designed to amplify the TA repeat (forward 5′-FITC-TAGACGCATGATATACTTCACCTAT-3′, reverse 5′-GCAGAATCAAATATCCAGATG-3′) and the CA repeat (forward 5′-GCCTAGTCAAATTCACAGAAAGCTA-3′, reverse 5′-HEX-TTGAAGATGGAGTGAGAAAAATACC-3′). The length of the 2 polymorphisms was determined by analysis of PCR products on 6% denaturing polyacrylamide gel8 with an ABI 377 DNA sequencer and the use of Genescan software (Applied Biosystems, Perkin-Elmer).
SSCP Analysis and Sequencing
A total of 33 primer pairs (please see online Table I, which can be accessed at http://atvb.ahajournals.org) were designed to evaluate the coding region, promoter A, and partial promoter B regions of the ER-α gene. Primers for the coding region flank the exons and allow for analysis of the adjacent splice junctions. Single-strand conformation polymorphism (SSCP) was performed from 60 subjects as described by Spinardi et al.20 Electrophoresis was performed on a 10% to 20% gradient polyacrylamide gel (ATTO Corp) and run at 100 V constantly for 10 to 15 hours in 0.5× TBE buffer at 3 different conditions: (1) at 4°C without glycerol and (2) at room temperature with 5% glycerol. Fragments that showed a variant by SSCP were reamplified and sequenced with fluorescently labeled dideoxy terminators with the use of a Thermo Sequenase II Kit (Amersham Pharmacia Biotech Inc) on an ABI 310 genetic analyzer (Applied Biosystems, Perkin-Elmer).
ANOVA was used to evaluate the association between phenotypic characteristics and genotype groups, and the Student t test was used when 2 groups were compared. The frequency distribution of genotypes was compared by using cross-tabulation and standard χ2 tests. Linkage disequilibrium between the different polymorphisms was analyzed by the estimate of a haplotype algorithm (EM) by using Arlequin software.21 To compare observed genotype frequencies with those expected under the Hardy-Weinberg equilibrium, contingency tables were used, with standard χ2 tests. Logistic regression analysis was used to predict CAD from the genotype of polymorphisms, with the known risk factors as covariates. Statview 5.0 software was used.
Characteristics of Study Subjects
The mean ages for men and women were not significantly different between the CAD and non-CAD groups (Table 1). The CAD group had a higher prevalence of conventional coronary risk factors (hypertension and diabetes for men; smoking, hypertension, and diabetes for postmenopausal women). Consistent with a previous report involving the Japanese population,22,23⇓ B-variant polymorphism was not detected in the present study.
Single Nucleotide Polymorphisms in the ER-α Gene Detected by SSCP Analysis and Sequencing
We identified 10 single nucleotide polymorphisms (Figure), including 4 that had been previously described in breast cancer patients,11,24⇓ 3 novel in introns, and 3 novel in promoter B; all exonic single nucleotide polymorphisms were synonymous. We identified a restriction enzyme whose cleavage pattern was altered by each variant or used a mismatch technique allowing restriction fragment length polymorphism analysis (see online Table II, which can be accessed at http://atvb.ahajournals.org). The distribution of genotypes, T9/T10 in intron 3 and T5/T6 in intron 5, was detected by PCR-SSCP and sequencing. All the genotype frequencies (please see online Table III, which can be accessed at http://atvb.ahajournals.org) followed the Hardy-Weinberg equilibrium. A T/G transversion in promoter B (−1989 relative to the first transcribed nucleotide in a homozygous for 14 TA repeats)9 was newly detected in the present study. The polymorphism of −1989T/G was in tight linkage disequilibrium with the TA repeat, PvuII and XbaI (please see online Table IV, which can be accessed at http://atvb.ahajournals.org), and it showed a significant association with CAD in men (P=0.02) and in postmenopausal women (P<0.05), with a higher frequency of G/G genotype in the CAD group (P=0.03). There was no significant difference in the severity of CAD or serum lipid levels, although HDL cholesterol was lower in subjects with the G/G genotype than in those with the T/T genotype (0.99±0.25 versus 1.14±0.32 mmol/L, P<0.05). Of a novel polymorphism (G/A in intron 4), G/A genotype and A allele frequency were significantly higher in postmenopausal women with CAD (P=0.03 and P=0.04, respectively) but not in men (please see online Table III). A stepwise logistic regression analysis with age, sex, body mass index, HDL cholesterol, and LDL cholesterol as covariates showed that the G/G genotype of −1989T/G (odds ratio [OR] 4.5, 95% CI 1.0 to 19.5; P=0.04), hypertension (OR 2.4, 95% CI 1.2 to 7.7; P=0.03), and diabetes (OR 3.0, 95% CI 1.6 to 10.0; P=0.02) were independent risks for CAD but that smoking and the X1/X1 genotype were not. When all the above confounding factors but HDL cholesterol and LDL cholesterol were forced, the G/G genotype weakened the statistical significance (OR 3.2, 95% CI 0.8 to 12.3; P=0.09), whereas the X1/X1 genotype showed significant risk for CAD (OR 5.0, 95% CI 1.2 to 21.4; P=0.03).
Effect of PvuII and XbaI Polymorphisms on CAD
The distribution of genotypes was in Hardy-Weinberg equilibrium, and as previously reported,5,6,9,22⇓⇓⇓ PvuII and XbaI were tightly linked, so that a higher than expected number of P1/P1 individuals were X1/X1 homozygotes and a higher than expected number of P2/P2 individuals were X2/X2 homozygotes (df 4, χ2=160.33; P<0.0001). XbaI genotype distributions were significantly different between the CAD and non-CAD groups in postmenopausal women (P=0.001) and in men (P=0.02), with a higher frequency of X1/X1 genotype in the CAD group (P=0.02). Serum HDL cholesterol, apoA-I, and apoA-II levels were significantly lower in patients with the P1/P1 genotype than in those with the P2/P2 genotype (P=0.04, P=0.03, and P=0.02, respectively) and in patients with the X1/X1 genotype compared with those with the X2/X2 genotype (P<0.05, P=0.02, and P=0.03, respectively). When genotypes were analyzed separately by sex, there was still a significant difference in HDL cholesterol as well as apoA-I for both polymorphisms and a significant difference between XbaI genotypes and apoA-II (P=0.04) in postmenopausal women but not in men. In addition, there was a significant difference in the mean age at onset of menopause among the genotypes of PvuII and XbaI in postmenopausal women (please see online Table V, which can be accessed at http://atvb.ahajournals.org). When we combined the 2 polymorphisms, 3 major haplotype alleles were recognized: P2X2 (54.0%), P1X1 (23.9%), and P1X2 (20.4%), thus demonstrating that 2 copies of the P1X1 haplotype are present more frequently than 1 or no copy of the P1X1 haplotype in postmenopausal women and men with CAD (P=0.01 and P=0.03, respectively). Serum apoA-I and apoA-II levels were significantly lower in patients with 2 copies of the P1X1 haplotype than in those with 1 or no copy of the P1X1 haplotype (P=0.02 and P=0.04, respectively), whereas HDL cholesterol showed a tendency to be lower in patients with 2 copies of the P1X1 haplotype (P=0.07). When the P1X1 haplotype was analyzed separately by sex, there was still a significant difference for apoA-I and apoA-II in postmenopausal women (P=0.02 and P=0.04, respectively) but not in men. No significant difference between P1X1 haplotype and HDL cholesterol was found for men or postmenopausal women (please see online Tables VI and VII, which can be accessed at http://atvb.ahajournals.org).
Effect of Dinucleotide Repeat Polymorphisms on CAD
According to the bimodal distribution pattern of TA alleles with a low distribution of 16 or 17 repeat alleles (please see online Figure IA, which can be accessed at http://atvb. ahajournals.org), subjects were classified into 3 groups: (1) those who carried 2 alleles with >17 repeats (designated [1,1]), (2) those who carried 1 allele with >17 repeats (designated [1,0]), and (3) those who carried 2 alleles with ≤17 repeats (designated [0,0]). A statistically significant correlation between TA repeat allelic variants and CAD was observed, with TA >17 alleles showing a higher frequency in postmenopausal women with CAD (P=0.04) but not in men (Table 2). No significant association was observed between TA repeats and serum lipid levels in men or in postmenopausal women.
In the present study, we report for the first time the distribution of CA repeats in intron 5 (please see online Figure IB). According to the bimodal distribution pattern of CA alleles with a low distribution of 20 or 21 repeat alleles, subjects were classified into 3 groups: (1) those who carried 2 alleles with >21 repeats (designated [1,1]), (2) those who carried1 allele with >21 repeats (designated [1,0]), and (3) those who carried 2 alleles with ≤21 repeats (designated [0,0]). No significant association was observed between CA repeats and CAD (Table 2) or serum lipid levels in men or in postmenopausal women.
In the present study, we reported a polymorphism analysis of the ER-α gene in the coding region, promoter A region, and partial promoter B regions in 295 patients with heterozygous FH, providing evidence that common genetic polymorphisms within the ER-α gene are associated with the risk of CAD in postmenopausal women and in men. The polymorphisms of −1989T/G in promoter B and XbaI in intron 1 are associated with an increasing risk of CAD in postmenopausal women and in men with heterozygous FH. Our results suggest that the TA repeat might be associated with CAD in postmenopausal women. The present study suggests that common allelic variants of the ER-α gene, existing in FH patients, may cause differential responsiveness to estrogen, although further experiments will be required to define the mechanism.
We first reported 3 novel polymorphisms in promoter B and found a high degree of linkage disequilibrium between −1989T/G and the variable length of TA repeats, −3012A/G, −1377T/C, PvuII, XbaI, 30T/C, and T9/T10. Subjects with the G/G genotype showed a significantly higher mean number of TA repeats and higher frequencies of −3012A/G, −1377C/C, P1/P1, X1/X1, 30C/C, and T10/T10 genotypes (P<0.0001). No comparable data are actually reported for other populations, although the linkage disequilibrium between the TA repeat and the polymorphisms of PvuII and XbaI was previously reported in white postmenopausal women.6,9⇓ The discovery of such a high degree of linkage disequilibrium in a group of 295 heterozygous FH individuals may have important implications and may partly explain previous discrepancies among ER-α polymorphism studies.5,6,9,10,22⇓⇓⇓⇓ The results from the present study indicate that there is a relationship between −1989T/G or its linked polymorphisms at the ER-α gene locus and CAD. However, the molecular mechanism by which −1989T/G or its linked polymorphisms is associated with CAD remains unclear. At least 3 different promoters have been identified in this gene.9,25–28⇓⇓⇓⇓ The first characterized promoter at the 5′ end of exon 1 was termed promoter A. It contains a TATA box and a CAAT element and possesses a single site of transcription initiation at +1.25 Subsequently, sequencing of upstream genomic DNA revealed a region at position −1.9 kb with an additional exon, denoted exon 1′, and an additional promoter, denoted promoter B.26 Several sites of transcription initiation are identified in this promoter.26,27⇓ There are few putative TATA box elements and no putative CAAT elements located near the proposed initiation sites, whereas a number of initiation response (INR) elements are present.28 Therefore, it is likely that transcription initiation is positioned via INR elements at this promoter. There are INR elements either at or just upstream from 3 of the 4 transcription initiation sites proposed by Grandien.27 There are several specificity protein 1 sites, 1 activator protein 1 site, and several estrogen response element (ERE)-like elements in the vicinity of exon 1′, which may augment transcription from promoter B.28 Promoter C, located >21 kb upstream from promoter A, has been postulated recently.28 Transcription of the ER-α gene from these promoters yields 3 different mRNA isoforms with unique 5′ untranslated regions but identical coding regions. Several studies have identified that promoter A and B are used in some breast cancer cell lines and endometrium but not in liver.27,28⇓ Expression of promoter C mRNA has been predominantly detected in liver, whereas only the promoter B mRNA expression has been detected in bone cells.27,28⇓ The factors controlling the level of expression of ER-α are not well characterized; however, these studies suggest that cell- and tissue-specific expression may be regulated by differential promoter usage. Recent studies have demonstrated that human vascular smooth muscle cells (VSMCs) express ER-α mRNA and protein and that ER-α in human VSMCs is capable of estrogen-dependent gene activation.29,30⇓ However, no study has ever investigated the expression of different promoter-mRNA isoforms in VSMCs. As detected in the present study, −1989T/G is located 2 bp downstream from 1 of the 4 identified transcription initiation sites in the promoter B reported by Grandien,27 which has several INR elements and Sp1 sites, 1 AP1 site, and 1 palindromic ERE-like element just upstream from this polymorphism.27,28⇓ Moreover, the TA repeat between promoter A and B regions, which was reported to be a predictor of the risk in postmenopausal osteoporosis,9 is ≈0.8 kb downstream from −1989T/G, and a high degree of linkage disequilibrium was observed between the 2 polymorphisms. We speculate that −1989T/G, or its linked TA repeats or perhaps yet unidentified polymorphisms within the linkage disequilibrium region, may directly or indirectly, by affecting promoter usage in VSMCs, influence the expression of the ER-α gene through transcriptional and translational regulation. Further study is required to identify whether or not the different modulation of ER-α gene expression that occurs at the transcription or translation level in human VSMCs associates with a polymorphism of −1989T/G or its linked polymorphisms.
The atheroprotective effects of estrogens were attributed to direct effects on the vessel wall and indirect effects on lipoprotein metabolism.31,32⇓ It is believed that estrogen leads to an increase in the mRNA for the LDL receptor.33 Many studies, including 1 large, randomized, controlled trial,32 have documented that estrogens decrease LDL cholesterol levels and increase HDL cholesterol levels. And it is reported that estrogens increase apoA-I production in hepatic cells by increasing the transcription of the apoA-I gene.34 A current study in 102 healthy Japanese school children has suggested that XbaI polymorphism might be related to LDL metabolism.35 Our results have shown that XbaI or its linked polymorphisms are associated with HDL and its principal apolipoproteins (apoA-I and apoA-II) in patients with heterozygous FH, although no significant association with LDL was found. This finding indicates that the ER-α gene might play a role in the transcriptional regulation of HDL metabolism in this special group of patients, whose primary defect is a mutation in the LDL receptor gene. There may be a true effect of XbaI or its linked polymorphisms on CAD, partially through an indirect effect via HDL metabolism and through a direct effect on coronary arteries. ER-α gene polymorphisms in HDL metabolism appear to be sex specific, but the number of subjects studied was too small for a definitive conclusion.
Our findings contradict those reported by Matsubara et al,22 who found no association between CAD and either of the 2 polymorphisms (PvuII and XbaI). However, in the present study, the X1 allele (25% versus 17%) and the X1/X1 genotype (9% versus 2%) were more frequent than those investigated in their study. The controversy may be partially explained by differences in the populations (FH patients versus the general population) or in the mean ages of CAD onset (50 versus 58 in men, and 61 versus 65 in postmenopausal women). It was reported that polymorphisms were not associated with serum lipid levels, including HDL cholesterol levels in their study. However, their study did not investigate apoA-I. Weel et al36 reported that homozygous P1/P1 women had a significantly earlier onset of menopause and a higher risk of surgical menopause than did homozygous P2/P2 women. We also found an earlier onset of menopause for women who carried the P1/P1 genotype or the X1/X1 genotype. Thus, the P1 or X1 allele appeared to be associated with an estrogen deficiency phenotype. We then speculated that the XbaI polymorphism might be a genetic risk factor for the age of CAD onset in women. However, in the study of Matsubara et al, the relationship between the 2 polymorphisms and the onset of menopause was not investigated. Sample size in the present study was calculated a priori, considering an α error <0.05 and a β error <0.1. This suggests that our sample size was large enough to minimize type I error and to assume a standard type II error for the XbaI polymorphism but not for PvuII. The statistical power with only 181 subjects in their study might have been too low to conclude that there was no association of XbaI or PvuII polymorphisms with CAD.
Interestingly, the studies in the Japanese population showed a striking difference in the genotype distribution from white populations. B-variant polymorphism was not detected among the Japanese population, although its allele frequency in white women has been reported to be 10% to 20%.23 The distribution of the X1/X1 genotype, which is likely associated with the estrogen deficiency phenotype in the Japanese population, was significantly lower than that observed in the white populations.5,6,9,22⇓⇓⇓ The relationship between these polymorphisms and CAD has not been widely studied in larger samples, and conflicting results have been reported in similar studies10,22⇓ and in studies of osteoporosis.5,6,9⇓⇓ The results we obtained in the present analysis clearly indicate the need for analyzing larger population samples in non-FH populations before reaching final conclusions.
In conclusion, the present study suggests that the polymorphisms in the ER-α gene, −1989T/G or its linked polymorphism XbaI, may confer risk for CAD in men and in postmenopausal women with FH. Carriers of the G/G genotype or its linked polymorphisms may benefit less from the cardiovascular protective effect of ER-α by its effect on HDL metabolism as well as by direct effects on blood vessels. Further studies in a non-FH population and the discovery of a functional molecular mechanism are necessary before −1989T/G might prove useful in the prediction of CAD.
This work has been supported by a scientific research grant from the Ministry of Education, Science, and Culture of Japan to H. Mabuchi (No. 09307010). We also thank Sachio Yamamoto and Mihoko Mizuno for technical assistance.
Received August 21, 2001; revision accepted February 13, 2002.
- ↵Barrett-Connor E. Sex differences in coronary heart disease: why are women so superior?: the 1995 Ancel Keys Lecture. Circulation. 1997; 95: 252–264.
- ↵Yaich L, Dupont WD, Cavener DR, Parl FF. Analysis of the Pvu II restriction fragment-length polymorphism and exon structure of the estrogen receptor gene in the breast cancer and peripheral blood. Cancer Res. 1992; 52: 77–83.
- ↵Lehrer S, Rabin J, Kalir T, Schachter BS. Estrogen receptor variant and hypertension in woman. Hypertension. 1993; 21: 439–441.
- ↵Becherini L, Gennari L, Masi L, Mansani R, Massart F, Morelli A, Falchetti A, Gonnelli S, Fiorelli G, Tanini A, Brandi ML. Evidence of a linkage disequilibrium between polymorphisms in the human estrogen receptorα gene and their relationship to bone mass variation in postmenopausal Italian women. Hum Mol Genet. 2000; 9: 2043–2050.
- ↵Kunnas TA, Laippala P, Penttila A, Lehtimaki T, Karhunen PJ. Association of polymorphism of human α oestrogen receptor gene with coronary artery disease in men: a necropsy study. BMJ. 2000; 321: 273–274.
- ↵Sudhir K, Chou TM, Chatterjee K, Smith EP, Williams TC, Kane JP, Malloy MJ, Korach KS, Rubanyi GM. Premature coronary artery disease associated with a disruptive mutation in the estrogen receptor gene in a man. Circulation. 1997; 96: 3774–3777.
- ↵Stone NJ, Levy RI, Fredrickson DS, Verter J. Coronary artery disease in 116 kindreds with familial type II hyperlipoproteinemia. Circulation. 1974; 49: 476–488.
- ↵Goldstein JL, Hobbs HH, Brown MS. Familial hypercholesterolemia.In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The Metabolic and Molecular Bases of Inherited Diseases. 8th ed. New York, NY: McGraw-Hill; 2001: 2863–2913.
- ↵Mabuchi H, Koizumi J, Shimizu M, Takeda R. Development of coronary heart disease in familial hypercholesterolemia. Circulation. 1989; 79: 225–232.
- ↵Spinardi L, Mazars R, Theillet C. Protocols for an improved detection of point mutations by SSCP. Nucleic Acids Res. 1991; 19: 4009.
- ↵Schneider S, Kueffer JM, Roessli D, Excoffier L. Arlequin Version 2.0: A Software for Population Genetics Data Analysis. Geneva, Switzerland: Genetics and Biometry Laboratory, University of Geneva; 2000.
- ↵Matsubara Y, Murata M, Kawano K, Zama T, Aoki N, Yoshino H, Watanabe G, Ishikawa K, Ikeda Y. Genotype distribution of estrogen receptor polymorphisms in men and postmenopausal women from healthy and coronary populations and its relation to serum lipid levels. Arterioscler Thromb Vasc Biol. 1997; 17: 3006–3012.
- ↵Zhang QX, Borg A, Wolf DM, Oesterreich S, Fuqua SA. An estrogen receptor mutant with strong hormone-independent activity from a metastatic breast cancer. Cancer Res. 1997; 57: 1244–1249.
- ↵Karas RH, Patterson BL, Mendelsohn ME. Human vascular smooth muscle cells contain functional estrogen receptor. Circulation. 1994; 89: 1943–1950.
- ↵Losordo DW, Kearney M, Kim EA, Jekanowski J, Isner JM. Variable expression of the estrogen receptor in normal and atherosclerotic coronary arteries of premenopausal women. Circulation. 1994; 89: 1501–1510.
- ↵Farhat MY, Lavigne MC, Ramwell PW. The vascular protective effects of estrogen. FASEB J. 1996; 10: 615–624.
- ↵Patrick TS, Yamamoto T, Goldstein JL, Brown MS. Increased mRNA for low density lipoprotein receptor in livers of rabbits treated with 17-ethinyl estradiol. Proc Natl Acad Sci U S A. 1986; 83: 792–796.
- ↵Lamon-Fava S, Ordovas JM, Schaefer EJ. Estrogen increases apolipoprotein (apo) A-I secretion in Hep G2 Cells by modulating transcription of the apo A-I gene promoter. Arterioscler Thromb Vasc Biol. 1999; 19: 2960–2965.