Estrogen and the estrogen receptor (ER) play a central role in bone metabolism as illustrated by the loss of bone mass after menopause and the osteopenia in individuals with defect aromatase or ER. We therefore wanted to investigate the effect of polymorphisms in the ER-α gene on bone mass, bone turnover, and the prevalence of osteoporotic fractures in a study of 160 women and 30 men with vertebral fractures and 124 women and 64 men who are normal. Three previously described polymorphisms, G261-C in exon 1 and T-C and A-G in intron 1, in the ER gene were determined by restriction fragment length polymorphism (RFLP) using BstUI, Pvu II, and Xba I after polymerase chain reaction (PCR). A TA repeat polymorphism in the promoter region was examined by PCR and electrophoresis. The distribution of BstUI, Pvu II, and Xba I RFLPs was similar in the osteoporotic patients and the normal controls. No significant differences could be shown in bone mass or bone turnover between the genotypes. The mean number of TA repeats was lower in patients with osteoporotic fractures, 17.3 ± 2.8 versus 18.6 ± 2.8 in the normal controls (p < 0.01). This also was reflected in a significantly increased odds ratio of osteoporotic fractures in individuals with 11–18 repeats of 2.64 (95% CIs, 1.61-4.34). Furthermore, bone mineral density (BMD) of the lumbar spine was lower in individuals with low mean number of repeats than in individuals with high mean number of repeats (0.790 ± 0.184 g/cm2 vs. 0.843 ± 0.191 g/cm2; p < 0.05). This difference also was found in BMD of the total hip. Using multiple linear regression, mean number of TA repeats was a predictor of lumbar spine BMD (p < 0.05) and a BMD-independent predictor of fractures (p < 0.05). Mean number of TA repeats was not associated with levels of biochemical markers of bone turnover. All four polymorphisms were in linkage disequilibrium. A TA repeat polymorphism in the ER gene is associated with increased risk of osteoporotic fractures and a modest reduction in bone mass. Polymorphisms in the first exon and first intron of the ER gene are not associated with osteoporotic fractures, bone mass, or bone turnover.
THE INCIDENCE of osteoporotic fractures is increasing. Therefore, prevention must be intensified. To increase the impact of prevention, a better detection of individuals at risk is needed. Osteoporosis is characterized by a combination of low bone mass and deteriorated microarchitecture. Bone mass is determined by the interaction of genetic, metabolic, and environmental factors. Genetic factors have been shown to be responsible for 40–75% of the interindividual variation.(1) Many candidate genes have been suggested and examined, transforming growth factor β1,(2) collagen Iα1,(3) vitamin D receptor,(4) interleukin-6,(5) interleukin-1 receptor antagonist,(6,7) and calcitonin receptor.(8) However, these polymorphisms or sequence variations are far from explaining all cases of osteoporosis.
Estrogen and estrogen receptors (ERs) are important for bone metabolism. Initiation of adolescent growth and bone acquisition requires estrogen.(9,10) Khosla et al.(11) showed that serum level of free estradiol is an important predictor of bone mass in both men and women. Furthermore, in older women, it has been shown that serum levels of estradiol are associated with osteoporotic fractures, bone mineral density (BMD), and rate of bone loss.(12–14) Osteoblasts, osteoclasts, and osteocytes express ERs.(15,16) Lack of estrogen results in increased osteoclast number and bone resorbing activity.(17,18)
Two different ER genes have been shown. The ER-α gene is located on chromosome 6q25-27, comprises eight exons, and spans more than 140 kilobases (kb).(19) The ER-β gene is located on chromosome 14q22-24, comprises eight exons, and spans approximately 40 kb. The two genes show 47% identity.(20)
In the ER-α gene a TA repeat polymorphism has been shown in the promoter.(21) The polymorphism is located 1174 base pairs (bp) upstream from the first exon. This polymorphism has been found to be associated with familiar premature ovarian failure(22) and bone mass.(23)
In codon 87 in the first exon a G to C polymorphism has been shown G261-C. This is a silent polymorphism, because the amino acid encoded by codon 87 is alanine in both cases.(24) In human mammary tumors the polymorphism affects the binding of estrogen to the receptor.(25) This polymorphism has been found to be associated with recurrent abortions and breast cancer in some studies.(26,27) Other studies could not confirm these findings.(28–30)
In the first intron two polymorphisms have been shown, T to C approximately 400 bp upstream from exon 2(31) and A to G approximately 350 bp upstream from exon 2. Alone or in combination, these polymorphisms have in some studies been shown to be associated with bone mass in men and women, postmenopausal bone loss, and response to hormone replacement therapy (HRT).(32–38) However, other studies could not confirm these findings.(39–42)
We wanted to examine the effect of these polymorphisms, alone and in combination, on bone mass, bone turnover, and prevalence of osteoporotic fractures in men and women.
MATERIALS AND METHODS
The study was a case control study. The osteoporotic group consisted of 160 women (mean age, 64.9 ± 8.4 years; range, 33–79 years; pre-/postmenopausal, 10/150; mean Z score [BMDlumbar spine], −1.88 ± 1.06) and 30 men (mean age, 55.8 ± 11.0 years; range, 22–77 years; mean Z score [BMDlumbar spine], −2.36 ± 1.47) with primary spinal osteoporosis defined by the presence of at least one nontraumatic fracture of the spine, referred to the Department of Endocrinology, Aarhus University Hospital. The diagnosis of primary osteoporosis was made after extensive examination for secondary causes. Spinal fracture was defined as a 20% or more reduction of the anterior, central, or posterior height of a vertebra. The normal control group comprised 124 normal women (mean age, 47.6 ± 13.3 years; range, 21–79 years; pre-/postmenopausal, 68/56; mean Z score [BMDlumbar spine], 0.02 ± 1.17) and 73 normal men (mean age, 51.1 ± 15.7 years; range, 22–78 years; mean Z score [BMDlumbar spine], −0.17 ± 1.39) without diseases or medications that could influence bone mass and turnover. The normal controls were recruited from the local community by invitations posted at places of work, senior citizens clubs, schools, educational institutions, hospitals, and at general practitioners. All participants were whites of northern European descent. For comparison of genotype frequencies we selected age-matched subgroups from the osteoporotic and normal groups. The age-matched groups comprised 80 osteoporotic women (mean age, 58.2 ± 6.4 years; range, 33–65 years), 80 normal women (mean age, 56.1 ± 7.8 years; range, 46–79 years), 30 osteoporotic men (mean age, 55.7 ± 11.0 years; range, 28–78 years), and 73 normal men (mean age, 51.1 ± 15.7 years; range, 22–78 years).
The study was approved by the local ethical committee and was conducted according to the Helsinki Declaration II.
Bone mass measurements
BMD of the lumbar spine and the following standard sites at the hip: femoral neck, trochanter, intertrochanteric region, and Wards triangle were assessed using dual-energy X-ray absorptiometry (DEXA) on a Hologic 1000 (Hologic Europe, Zaventum, Belgium) or a Norland (Norland Corp., Gammatec AS, Værløse, Denmark) bone densiotometer. Results obtained on the Norland densiotometer were corrected for the difference between the two densiotometers using the correction formulas suggested by Genant et al.(43) All BMD values were adjusted for age and gender.
Biochemical markers of bone turnover
Serum samples were collected in the morning after an 8-h fasting period. All samples from osteoporotic patients were collected before institution of antiosteoporotic treatment. Urine samples were 24-h samples.
Serum cross-linked carboxy-terminal telopeptide of type I collagen (s-ICTP) was measured by an equilibrium radioimmunoassay, the intra-assay CV was 5% and the interassay CV was 6%.(44) S-osteocalcin was determined by a radioimmunoassay using rabbit antiserum against bovine bone-gla-protein.(45) The intra-assay CV was 5% and the interassay CV was 10%.
TA repeat polymorphism in the promoter
The TA repeat polymorphism was examined by a polymerase chain reaction (PCR)-based method as previously described.(23) In short, PCR was performed in a final volume of 25 μl containing 100 ng genomic DNA, 150 ng of each primer and AmpliTaqGold polymerase (Perkin-Elmer, Allerød, Denmark) using standard conditions on a Perkin-Elmer Termocycler 2400: 30 cycles of 95°C for 1 minute, 59°C for 1 minute, and 74°C for 1 minute. Before the first cycle, initial denaturation was performed at 95°C for 10 minutes, and the last cycle was followed by an extension step of 8 minutes at 74°C. The downstream primer is 5′ GAC GCA TGA TAT ACT TCA CC and the upstream primer is 5′ GCA GAA TCA AAT ATC CAG ATG. The PCR product was analyzed on a 4.25% polyacrylamide gel electrophoresis (PAGE) gel using the GeneScan software package (Perkin-Elmer). The TA repeat polymorphism was examined in the age-matched subgroups (80 osteoporotic women and 80 normal women; 30 osteoporotic men and 73 normal men).
Restriction site polymorphisms
The G261-C in the first exon was examined by a PCR-based method as previously described.(46) In short, PCRs were performed in a final volume of 25 μl containing 100 ng genomic DNA, 150 ng of each primer, and AmpliTaqGold DNA polymerase (Perkin-Elmer) using standard conditions on a Perkin-Elmer Termocycler 2400: 35 cycles of 95°C for 1 minute, 66°C for 1 minute, and 72°C for 1 minute. Before the first cycle, initial denaturation was performed at 95°C for 10 minutes, and the last cycle was followed by an extension step of 10 minutes at 72°C. The downstream primer is 5′ CGC GCA GGT CTA CGG TCA G and the upstream primer is 5′ GCT GCG GCG GCG GGT GCA. The PCR products were digested for 1 h at 60°C with BstUI (New England Biolabs, Medinova Scientific AS, Hellerup, Denmark) according to the manufacturers instructions and were analyzed on a 2% agarose gel. Absence and presence of the BstUI restriction site are specified as B and b, respectively.
The polymorphic Pvu II and Xba I restriction enzyme sites in intron 1 were examined by a PCR-based method as previously described.(47) In short, PCRs were performed in a final volume of 50 μl containing 200 ng genomic DNA, 300 ng of each primer, and AmpliTaqGold DNA polymerase (Perkin-Elmer) using standard conditions on a Perkin-Elmer Termocycler 2400: 35 cycles of 95°C for 30 s, 61°C for 40 s, and 72°C for 90 s. Before the first cycle, initial denaturation was performed at 95°C for 10 minutes, and the last cycle was followed by an extension step of 6 minutes at 72°C. The downstream primer is 5′ CTG CCA CCC TAT CTG TAT CTT TTC CTA TTC TCC 3′ and the upstream primer is 5′ TCT TTC TCT GCC ACC CTG GCG TCG ATT ATC TGA 3′. The PCR products were digested overnight with Pvu II (New England Biolabs) and Xba I (New England Biolabs) at 37°C and according to the manufacturers instructions and were analyzed on 2% agarose gels. Absence and presence of the Pvu II and Xba I restriction sites are specified as P and p and X and x, respectively.
Differences in prevalence of the genotypes between osteoporotic patients and age-matched normal controls were tested using χ2 test. The effect of genotype and haplotypes (combined genotypes) on BMD and levels of biochemical markers was evaluated by analysis of variance (ANOVA). Differences in BMD and levels of biochemical markers between groups were tested using Student's t-test for unpaired data. Linkage between the different polymorphisms was examined by χ2 test of actual and expected distribution of haplotypes. Expected prevalence was calculated from the prevalence of each genotype. Prevalence (AABB) = prevalence (AA) × prevalence (BB). The level of significance was set at 0.05.
Distribution of BstUI (G261-C), Pvu II, and Xba I restriction site genotypes in osteoporotic patients and normal controls is shown in Table 1. The allele frequencies of the three polymorphisms are in Hardy-Weinberg equilibrium (χ2 = 0.30, p = 0.86; χ2 = 0.24, p = 0.89; and χ2 = 0.10, p = 0.95, respectively). There was no difference between osteoporotic patients and normal controls in prevalence of these genotypes.
Table Table 1.. Distribution of BstUI, Pvu II, and Xba I Genotypes in Osteoporotic Patients and Normal Controls
BMD of the lumbar spine was 0.834 ± 0.167 g/cm2 in subjects with the BB genotype, 0.839 ± 0.150 g/cm2 in subjects with the Bb genotype, and 0.886 ± 0.373 g/cm2 in subjects with the bb genotype (NS). Comparing BMD of the femoral neck and the total hip between the three genotypes revealed a similar result (Table 2). Only 3 individuals exhibited the bb genotype; they had higher levels of s-osteocalcin (25.6 ± 10.3 μg/liter vs. 16.4 ± 7.1 μg/liter and 16.4 ± 6.3 μg/liter in individuals with the BB and Bb genotype, respectively; p < 0.05) and s-ICTP (4.7 ± 0.8 μg/liter vs. 3.2 ± 1.0 μg/liter and 3.1 ± 1.4 μg/liter in individuals with the BB and Bb genotype, respectively; p < 0.05; Table 2).
Table Table 2.. Age, Height, Weight, Age- and Sex-Corrected BMD, and Biochemical Markers of Bone Turnover in BstUI, Xba I, and Pvu II Genotypes (Mean ± SD)
BMD of the lumbar spine was 0.861 ± 0.166 g/cm2 in subjects with the XX genotype, 0.828 ± 0.164 g/cm2 in subjects with the Xx genotype, and 0.835 ± 0.167 g/cm2 in subjects with the xx genotype (NS). BMD of the femoral neck and the total hip was not different between the genotypes. No differences in s-osteocalcin and s-ICTP could be shown (Table 2).
BMD of the lumbar spine was 0.844 ± 0.176 g/cm2 in individuals with the PP genotype, 0.831 ± 0.161 g/cm2 in individuals with the Pp genotype, and 0.834 ± 0.165 g/cm2 in individuals with the pp genotype (NS). BMD of the femoral neck and the total hip was comparable between individuals with the PP, Pp, and pp genotypes. Furthermore, no differences could be found between genotypes in biochemical markers of bone turnover (Table 2).
Genotypes of the three polymorphisms were combined two and two and all three together. There was no difference in prevalence of these combined genotypes or haplotypes between osteoporotic patients and normal controls (Tables 3 and 4). The haplotypes XxPP, xxPP, xxPp, BBxxPP, BbXXPP, bbXXPP, bbXxPP, bbXxPp, bbxxPp, and bbxxpp were not found in our material. Furthermore, no differences could be shown in BMD or biochemical markers of bone turnover between these combined genotypes (data not shown). The distribution of the combined genotypes (haplotypes) is not random (χ2 = 16.6 − 175; p = 0.03 − 1.19 × 10−33); the polymorphisms are in linkage disequilibrium. When all three genotypes were combined, a similar result was obtained (χ2 = 133; p = 5.18 × 10−21).
Table Table 3.. Distribution of Xba I and Pvu II Combined Genotypes in Osteoporotic Patients and Normal Controls
Table Table 4.. Distribution of BstUI, Xba I, and Pvu II Combined Genotypes in Osteoporotic Patients and Normal Controls
The number of TA repeats varied between 8 and 26, mean number of repeats in the participants ranged from 11 to 25.5, and the distribution was bimodal (Fig. 1). Mean number of repeats was lower in osteoporotic patients than in normal controls (17.3 ± 2.8 vs. 18.6 ± 2.8; p < 0.01). This also was reflected in a significantly increased odds ratio of osteoporotic fractures in individuals with 11–14 repeats (lowest quartile of mean number of repeats) of 2.48 (95% CIs, 1.09-5.67) and in individuals with 11–18 (lowest half of mean number of repeat) of 2.64 (95% CIs, 1.61-4.34). The difference in mean number of repeats also was found in women (16.9 ± 2.9 in osteoporotic women vs. 18.8 ± 3.0 in normal women; p < 0.01) but not between osteoporotic and normal men (17.9 ± 2.4 vs. 18.4 ± 2.7, NS).
We grouped the genotypes in two groups of equal numbers of participants: group 1 (low mean number of TA repeats), 11–18.5; and group 2 (high number of repeats), 19–25.5. BMD of the lumbar spine was lower in individuals with low mean number of repeats than in individuals with high mean number of repeats (0.790 ± 0.184 g/cm2 vs. 0.843 ± 0.191 g/cm2; p < 0.05; Fig. 2). The individuals with low mean number of repeats also had low BMD at the femoral neck and the total hip. The difference was only significant for BMD of the total hip (p < 0.05). No significant effect of the TA repeat polymorphism could be shown on biochemical markers of bone turnover, height, or weight. Multiple linear regression analysis revealed that mean number of TA repeats was a predictor of lumbar spine BMD (p < 0.05) and osteoporotic fractures (p < 0.005). If lumbar spine BMD was included in the analysis, mean number of TA repeats was found to be a BMD-independent predictor of fractures (p < 0.05).
The TA repeat polymorphism is linked to the single nucleotide polymorphisms in exon and intron 1. Mean number of TA repeats is lowest in individuals with the BB, xx, or pp genotypes and highest in individuals with bb, XX, or PP genotypes (ANOVA, p < 0.0001; Fig. 3).
Estrogen is known to play an important role in bone metabolism in both women and men, and the ER gene therefore is an obvious candidate gene for susceptibility to development of osteoporosis. In this study, we have shown that a TA repeat polymorphism located in the promoter 1174 bp upstream from exon 1 in the ER-α gene affects bone metabolism. Low mean number of TA repeats was associated with a modest reduction in bone mass both at the lumbar spine and at the hip and, most importantly, with risk of osteoporotic fractures. This TA repeat polymorphism in the promoter could be responsible for the reduced bone mass and increased fracture risk through changes in ER number caused by changes in messenger RNA (mRNA) production or stability. Piva et al.(48) found no association between the amount of mRNA produced in malignant cell lines and number of TA repeats; however, DNA from these cell lines showed a low degree of heterozygosity, suggesting loss of somatic alleles. Previously, a low number of repeats has been found to be associated with familiar premature ovarian failure(22) and if this polymorphism affects age at menopause, this could be part of the explanation for the lower bone mass and increased fracture risk.(49) Furthermore, this would explain why no association was found in men. Further studies will be needed to clarify the mechanism behind the associations.
In this study we did not find any effects of the Xba I and Pvu II polymorphisms or the derived haplotypes on bone mass or fracture risk. So far, only a few studies have indicated associations between the Xba I and Pvu II polymorphisms and bone mass. Ongphiphadhanakul et al.(36) found that the PP genotype is associated with high bone mass in premenopausal Thai women but not in postmenopausal women. In American premenopausal women, Willing et al.(34) showed that the pp and xx genotypes were associated with reduced BMD but not with changes in BMD over 3 years. In Japanese women, the xx genotype was associated with low bone mass only in premenopausal women.(35) Only one study by Oi et al. has reported association between the pp genotype and low bone mass in postmenopausal women and men.(32) However, a Japanese study did not find associations between the genotypes and bone mass but found that the haplotype PPxx was associated with low bone mass.(33) Several other studies,(39,41,42,50) mainly including postmenopausal women, did not find any association between these intronic polymorphisms and bone mass. Only one study has examined the effect of these polymorphisms on fracture risk. Vandevyver et al.(50) found no association between the Pvu II polymorphism and risk of hip, spine, or forearm fractures. In the present study, we only examined the risk of vertebral fractures and found no association with the intronic polymorphisms. If these polymorphisms have an effect on bone mass in premenopausal women, this effect seems to be lost during menopause and aging and is not reflected in postmenopausal bone mass or risk of osteoporotic fractures. Furthermore, we did not find any effect of these polymorphisms on bone mass or fracture risk in men.
No functional studies have been performed to clarify if and how these intronic polymorphisms could affect the ER. Theoretically, polymorphisms in introns could affect mRNA production because introns and especially intron 1 have been shown to contain regulatory sequences such as enhancers of mRNA production(51,52) and that polymorphisms within the first intron can have significant effect on the level of protein synthesis.(53) However, because the results of association studies have been conflicting, a more likely explanation for the associations shown is that the Pvu II and Xba I polymorphisms are in linkage disequilibrium with other polymorphisms in the promoter or coding region of the gene or with other nearby genes like the insulin-like growth factor I gene or the parathyroid hormone-related peptide gene, which are located on the same chromosome as ER-α. We have shown in this study that all four examined polymorphisms are strongly linked. If the TA repeat polymorphism is the pathogenic polymorphism, the associations previously shown between the intronic polymorphisms and bone mass can be explained through linkage with the TA repeat polymorphism.
The haplotypes XXPP, XxPp, and xxpp are the most frequent in our population and XXpp and xxPP are rare. It is therefore most likely that X and P are linked and X and p are not linked. Furthermore, high mean number of TA repeats is linked with the XX and PP genotypes of the Xba I and Pvu II polymorphisms. We only found 4 normal individuals and no osteoporotic patients with the PPxx genotype. This genotype is rare (2.6%) in our population and is not associated with reduced bone mass or osteoporotic fractures as previously reported by Kobayashi et al.(33)
The G261-C polymorphism in exon 1 has been examined for effects on the ER. Normally, the produced amount of ER mRNA is correlated with the binding capacity of estrogen; however, breast cancer tissue with the bb genotype displayed reduced binding capacity despite normal mRNA levels.(25) Encouraged by this finding, the polymorphism has been examined in many different diseases in which estrogen is known to play a role. In women with ER-positive breast cancer, history of spontaneous abortions was associated with presence of the b allele of the ER. This association could not be shown in women with ER-negative breast cancer or women without breast cancer.(27–29) Furthermore, no associations have been shown between G261-C and incidence of breast cancer.(30,46) Testis cancer is related to a high level of estrogen during fetal life; however, Heimdal et al.(54) did not find testis cancer to be associated with this polymorphism. In a Japanese study comprising 87 patients with coronary artery disease and 94 healthy controls, no associations could be shown between the BstUI, Xba I, and Pvu II polymorphisms in the ER gene and prevalence of coronary artery disease or serum lipid levels.(55) Now, we have shown that bone mass and risk of osteoporotic fractures are not associated with the BstUI polymorphism in women and men.
This study has certain limitations. We have only included osteoporotic patients with vertebral fractures and the increased fracture risk shown in individuals with low number of TA repeats is not necessarily true for forearm or hip fractures. The power of detecting a difference of 0.5 SD in BMD between genotypes was 87–97% and the power of detecting a 20% difference in allele frequencies was 87–95%. However, in subgroup analyses in women and men, the power was only 75% and 40–50%, respectively. Therefore, results obtained in these subgroups must be interpreted with caution.
In conclusion, we have shown that a TA repeat in the promoter of the ER gene is associated with increased risk of osteoporotic fractures and a modest reduction in bone mass, whereas polymorphisms in exon 1 and intron 1 are not. Further studies will be needed to clarify the mechanisms underlying the associations shown.
The authors thank The Novo Nordisk Fonden, The Institute of Experimental Clinical Research, University of Aarhus, The Fonden til Lægevidenskabens Fremme, and The Danish Center for Molecular Gerontology for financial support.