|
|
1 Gene to Bone Density and the Risk of Osteoporotic Fractures in Postmenopausal Women
| |||||||||||||||||||||||||||||||||||||||||||||
Background Osteoporosis is a common disorder with a strong genetic component. One way in which the genetic component could be expressed is through polymorphism of COLIA1, the gene for collagen type I
Methods We determined the COLIA1 genotypes SS, Ss, and ss in a population-based sample of 1778 postmenopausal women using a polymerase-chain-reaction–based assay. We then related the genotypes to bone mineral density and the occurrence of osteoporotic fractures in these women.
Results As compared with the 1194 women with the SS genotype, the 526 women with the Ss genotype had 2 percent lower bone mineral density at the femoral neck (P = 0.003) and the lumbar spine (P = 0.02); the 58 women with the ss genotype had reductions of 4 percent at the femoral neck (P = 0.05) and 6 percent at the lumbar spine (P = 0.005). These differences increased with age (P = 0.01 for modification by age of the effect of COLIA1 on femoral-neck bone density, and P = 0.004 for modification of the effect on lumbar-spine bone density). Women with the Ss and ss genotypes were overrepresented among the 111 women who had incident nonvertebral fractures (relative risk per copy of the s allele, 1.5; 95 percent confidence interval, 1.1 to 2.1).
Conclusions The COLIA1 polymorphism is associated with reduced bone density and predisposes women to osteoporotic fractures.
1, a bone-matrix protein.
,25-dihydroxyvitamin D–resistant rickets.8,9 Some, but not all, subsequent population-based and twin studies found that polymorphisms of the vitamin D receptor were associated with differences in bone density.10,11 A man with a loss-of-function mutation of the estrogen-receptor gene had osteoporosis,12 but intronic polymorphisms of this gene have not been consistently associated with differences in bone density in population-based studies of women.13,14
Similarly, girls with loss-of-function mutations in the aromatase P450 gene, who cannot synthesize estrogen, have osteoporosis,15 but no population-based studies of this gene have been published. In more recent pedigree studies, a locus for "high peak bone density" and the locus for the osteoporosis–pseudoglioma syndrome, a recessive disorder characterized by juvenile-onset blindness and osteoporosis, were mapped to the same region on chromosome 11,16,17 suggesting the presence of a gene with effects on bone density. Polymorphisms of other candidate genes, such as the genes for interleukin-6 and transforming growth factor 
, have also been associated with differences in bone density in some populations.18,19
The genes encoding collagen types I1 and I2 (COLIA1 and COLIA2, respectively) are also important candidates for the genetic regulation of bone density, because mutations that affect the coding regions of these genes cause osteogenesis imperfecta.20 Although the coding regions of collagen genes are normal in the vast majority of patients with osteoporosis,21 we described a novel guanine-to-thymidine polymorphism at the first base of a binding site for the transcription factor Sp1 in the first intron of the COLIA1 gene that was associated with low bone density and increased occurrence of osteoporotic vertebral fracture in 299 British women.22 To determine whether this polymorphism predicts bone density and the risk of osteoporotic fracture in other populations, we studied the relation of the COLIA1 polymorphism to bone density and osteoporotic fracture in 1778 postmenopausal women in the Netherlands.23
Methods
Study Subjects
The Rotterdam Study is a population-based cohort study of 7983 subjects who are at least 55 years old and who live in the Ommoord district of Rotterdam, the Netherlands. The study was designed to document the occurrence of disease in the elderly in relation to several potential determinants.23 A total of 10,275 persons, of whom 9161 (89 percent) were living independently, were invited to participate in the study in 1991. Among the subjects living independently, the overall response rate was 77 percent for the home interview and 71 percent for the examination in a research center, where anthropometric characteristics and bone mineral density were measured and blood samples were taken. The Rotterdam Study was approved by the medical ethics committee of the Erasmus University Medical School, and written informed consent was obtained from each subject.
The association between the COLIA1 genotype and osteoporosis was analyzed in a subgroup of women participating in the study. Base-line measurements of bone mineral density were available for 5931 subjects who were living independently. Of these, 1453 were excluded because they were more than 80 years of age, used a walking aid, had diabetes mellitus, or were taking diuretics, estrogen, thyroid hormone, or cytostatic drugs. From the 4478 remaining subjects, we studied a random sample of 2000 women who were 55 to 80 years old. Bone-density data or DNA samples were not available for 222 women, resulting in a final study group of 1778 women.
Measurements
Height and weight were measured at the initial examination, with the subject in a standing position without shoes. Bone mineral density (in grams per square centimeter) was determined by dual-energy x-ray absorptiometry (DPX-L densitometer, Lunar, Madison, Wis.) at the femoral neck and lumbar spine (vertebrae L2, L3, and L4), as described elsewhere.24 Dietary intake of calcium (in milligrams per day) during the preceding year was assessed by a food-frequency questionnaire and adjusted for energy intake. Age at menopause and current use of cigarettes and alcohol were assessed by questionnaire. For 1423 women (80 percent), lateral radiographs of the spine from the fourth thoracic to the fifth lumbar vertebra were obtained at base line and examined for the presence of prevalent vertebral fractures by morphometric analysis, as previously described.25 Incident nonvertebral fractures, including hip, wrist, and other fractures, were recorded, confirmed, and classified by a physician over a mean follow-up period of 3.8 years.
Determination of COLIA1 Genotypes
Genomic DNA was extracted from samples of peripheral venous blood according to standard procedures, and the intronic polymorphism of the COLIA1 gene was detected by the polymerase chain reaction (PCR) with a mismatched primer that introduces a diallelic restriction site, as previously described.22 The test discriminates two alleles, S and s, which correspond to the presence of guanine and thymidine, respectively, as the first bases in the Sp1-binding site in the first intron of the gene for COLIA1. The reaction mixture of 25 µl contained 100 ng of genomic DNA, 50 mM potassium chloride, 10 mM TRIS–hydrochloric acid (pH 8.3), 1.5 mM magnesium chloride, 0.2 mM dideoxynucleotide triphosphates, 150 ng of each primer, and 0.2 unit of Super Taq polymerase (HT Biotechnology, Cambridge, United Kingdom). The reactions were performed in a DNA thermocycler (mode 480, Perkin–Elmer, Foster City, Calif.) with a cycling protocol of 94°C, 60°C, and 72°C for one minute each, for 35 cycles. Ten microliters of PCR product was digested with 2 units of MscI restriction enzyme and 1.2 µl of a buffer, containing 150 mM TRIS–hydrochloric acid (pH 7.5), 250 mM sodium chloride, and 35 mM magnesium chloride, by incubation for 30 minutes at 37°C. The digestion products underwent electrophoresis on a 3 percent NuSieve agarose gel (FMC BioProducts, Rockland, Me.) in 44.5 mM TRIS, 44.5 mM boric acid, and 1 mM disodium EDTA for 300 volt-hours. The separation patterns were documented by Polaroid photography under ultraviolet illumination (302 nm). To confirm the accuracy of the genotyping, repeated analysis was performed on 400 randomly selected samples, including those from all ss homozygotes. No discrepancies were found.
Statistical Analysis
Bone mineral density and other relevant clinical variables of the three genotype groups were compared by analysis of covariance. Multiple linear regression was used to adjust bone-density values for confounding factors such as age, anthropometric variables, calcium intake, smoking, and alcohol use. A multivariate linear regression model was designed to test for the changing effects of COLIA1 genotype on bone density with age. For this analysis, the subjects were classified into five-year age groups (55 to 59, 60 to 64, 65 to 69, 70 to 74, and 75 to 80 years), and the COLIA1 genotypes were assigned values of 0, 1, or 2, according to the number of s alleles present. The analysis included an interaction term defined as age group multiplied by gene dose (the number of s alleles). The variance of bone density that could be explained by the COLIA1 genotype was calculated by linear regression analysis adjusted for weight and age. The chi-square test was used to test for Hardy–Weinberg equilibrium, the distribution of alleles in women of different age groups, and the distribution of alleles in women with and without fractures. Odds ratios (with 95 percent confidence intervals) were calculated by multivariate logistic-regression analysis to estimate the relative risk of osteoporotic fracture. All statistical tests were two-sided.
Results
Characteristics of the Study Subjects
Table 1 shows anthropometric, dietary, and bone-density measurements according to COLIA1 genotype. The allelic frequencies (82 percent for the S allele and 18 percent for the s allele) and the distribution of genotypes were similar to those reported in a previous study22 and were as predicted by the Hardy–Weinberg equation, confirming that no genotype was being selectively sampled (P = 1.00). The three genotype groups did not differ significantly in age, age at menopause, dietary calcium intake, smoking habits, alcohol intake, or height. There was a significant gene-dose effect on body weight (-0.97 kg per copy of the s allele, P = 0.03): body weight was lower in the Ss group than in the SS group, and lowest in the ss group.
|
The three genotype groups differed significantly in bone mineral density at the femoral neck and the lumbar spine. At both sites, bone mineral density was highest in the SS group and lowest in the ss group (Table 1). Bone mineral density in the Ss group was 2 percent less than that in the SS group at both sites (P =0.003 for femoral neck and P = 0.02 for lumbar spine). In the ss group, bone mineral density in the femoral neck was 4 percent less than that in the SS group (P = 0.05), and in the lumbar spine it was 6 percent less than that in the SS group (P = 0.005). Table 2 shows that these differences corresponded to significant gene-dose effects at the femoral neck and lumbar spine (P<0.001 for both sites). The effects of COLIA1 genotype on bone density of the Ward's triangle and trochanter sites at the hip were similar (data not shown). The percentage of the variance in bone density explained by the COLIA1 genotype was 0.3 percent at the femoral neck and 0.4 percent at the lumbar spine. When the bone-density values were adjusted for confounding variables resulting from the association between genotype and body weight, the differences between genotype groups were smaller, but they were still statistically significant.
|
Bone density declines with age, and genetic effects on bone density could theoretically be mediated by differences in the age-related rate of bone loss,26 as well as by differences in peak bone density. We therefore assessed the effect of age on the relation between COLIA1 genotype and bone density. When the study subjects were divided into five-year age categories, the differences in bone density between the genotype groups were small among the younger women (55 to 65 years old) but became larger with increasing age (Figure 1A and Figure 1B). Among women 75 to 80 years of age, bone mineral density in the Ss group was 5 percent lower at the femoral neck (P = 0.03) and 3 percent lower at the lumbar spine (P = 0.26) than that in the SS group. In the ss group, bone mineral density was 12 percent lower at the femoral neck (P = 0.04) and 20 percent lower at the lumbar spine (P = 0.004) than that in the SS group. These numbers indicate an increased gene-dose effect with increasing age at both the femoral neck and the lumbar spine (Table 2).
|
Association between Genotype and Fractures
Table 3 shows the distribution of incident nonvertebral fractures and prevalent vertebral compression fractures according to COLIA1 genotype. Significantly more women with the Ss genotype had incident nonvertebral fractures or prevalent vertebral fractures than women with the SS genotype. Among women with the ss genotype, this difference increased for incident nonvertebral fractures (P = 0.02) but not for prevalent vertebral fractures (P = 0.33).
|
Discussion
Genetic factors have an important role in determining bone density,2,3,4,5,6,27 and bone density is an important determinant of the risk of osteoporotic fracture.28,29 The genes for collagen types I1 and I2 (COLIA1 and COLIA2) are important candidates for regulators of bone density and risk of fracture, because point mutations that affect their coding regions result in osteogenesis imperfecta.20 Although similar mutations in the coding regions of the collagen genes have not been found in patients with osteoporosis,21 the polymorphism that we described, which affects a binding site for the transcription factor Sp1 in a regulatory region of the COLIA1 gene, was associated with differences in bone density and risk of fracture in 299 British women.22 The data presented here confirm and extend these findings to a population-based cohort of 1778 postmenopausal women from the Netherlands. We found a consistent association between this COLIA1 polymorphism and differences in bone density at the spine and hip, with evidence of gene-dose effects at both sites: bone-density values were highest in women with the SS genotype, intermediate in those with the Ss genotype, and lowest in those with the rare ss genotype.
These differences in bone density were accompanied by differences in body weight, and the difference between genotypes in bone density was reduced in magnitude after correction for weight. This is not unexpected in view of the close association between weight and bone density30,31 and suggests that the association between COLIA1 genotype and bone density may be mediated in part by a genetic effect of COLIA1 on body weight.
The genotype-specific differences in bone density were small in the study group as a whole but increased with age. The allelic frequencies did not change with age, an observation that argues against selection bias due to mortality-related effects of the polymorphism as an explanation for this effect. This raises the possibility that the COLIA1 polymorphism may be a marker for accelerated bone loss in older women, rather than a marker for lower peak bone density.
The overrepresentation of the Ss and ss genotypes among women with osteoporotic fractures was not unexpected, in view of the well-established relation between bone density and risk of fracture,28,29 which was also found in this study. The relative risk of fracture, however, was about two to three times as great as that predicted by the differences between genotypes in bone density, and it remained significant after adjustment for weight and bone density. This raises the possibility that the COLIA1 polymorphism may act as a marker for differences in bone quality, such as bone structure or matrix composition, as well as bone quantity. The number of fractures was small, however.
The mechanisms by which the different COLIA1 alleles affect bone density and fracture risk are not known. Oligonucleotides corresponding to the s allele bind the Sp1 protein with greater affinity than those corresponding to the S allele,32 and subjects who are heterozygous for the polymorphism have three times as many transcripts derived from the s allele as from the S allele (unpublished data). This raises the possibility that carriage of the s allele may be associated with a disturbance in the relative abundance of COLIA1 and COLIA2 messenger RNAs, as occurs in osteogenesis imperfecta.33
Studies of the genetic basis of osteoporosis are important to identify genes that act as regulators of bone density, bone quality, and the risk of fracture. Genetic markers for osteoporosis could be useful clinically for identifying subjects at risk of osteoporosis and for targeting preventive treatment. The association we describe between COLIA1 genotype and osteoporotic fracture in postmenopausal women, coupled with the divergence of bone-density values with increasing age between different genotype groups, raises the possibility that genotyping at the polymorphic COLIA1 Sp1 site may provide information on susceptibility to osteoporotic fracture that could complement information gained from bone-density measurements alone.
Supported by grants from the Dutch Praeventiefonds (002824890), the Nestor stimulation program for geriatric research in the Netherlands (Ministry of Health and Ministry of Education), the Municipality of Rotterdam, the Netherlands Organization for Scientific Research, the European Commission, the Chief Scientist's Office of the Scottish Home and Health Department, the Wellcome Trust, and the Grampian Osteoporosis Trust.
Source Information
From the Departments of Internal Medicine III (A.G.U., J.P. T.M.L., H.A.P.P.) and Epidemiology and Biostatistics (A.G.U., H.B., Q.H., F.Y., A.H., H.A.P.P.), Erasmus University Medical School, Rotterdam, the Netherlands; and the Department of Medicine and Therapeutics, University of Aberdeen, Aberdeen, United Kingdom (F.E.A.M., S.F.A.G., S.H.R.).
Address reprint requests to Dr. Uitterlinden at the Department of Internal Medicine III, FGG-EUR, Erasmus University Medical School, P.O. Box 1738, 3000 DR Rotterdam, the Netherlands.
References
,25-dihydroxyvitamin D-resistant rickets resulting from a mutation in the vitamin D receptor deoxyribonucleic acid-binding domain. J Clin Endocrinol Metab 1994;78:313-316. [Abstract]
1 gene. Nat Genet 1996;14:203-205. [CrossRef][Medline]
1 gene. (Ph.D. thesis. Aberdeen, Scotland: University of Aberdeen, 1996.)
| |||||||||||||||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
0.05, and the dagger P<0.01 for the comparison with the SS genotype.