Osteoporosis is characterized by reduced bone mineral density (BMD), deterioration of bone microarchitecture and increased risk of fractures (1). Genetic factors play an important role in the pathogenesis of low bone mass and compromised bone quality. One of the most important candidate genes predisposing to osteoporosis is the collagen type Iα1 (COLIA1) gene. It encodes the α1I protein chain of type I collagen – the major protein of bone. Mutations that affect the coding region of COLIA1 gene cause osteogenesis imperfecta (2). A single G→T nucleotide polymorphism might affect the binding site for the transcription factor Sp1 in the COLIA1 gene (3). The polymorphic Sp1 site lies within the first intron of the COLIA1 gene in the region regulating the collagen transcription (4). This particular polymorphism was associated in several studies with low BMD and an increased risk of osteoporotic fractures (4-9). There is evidence of allele-specific differences in the binding of the Sp1 protein to the polymorphic recognition site, in the allele-specific transcription, in the collagen production, and in the bone strength in patients with different genotypes (10). Oligonucleotides corresponding to the s allele bind the Sp1 protein with higher affinity than those corresponding to the S allele (10) and Ss heterozygotes have three times more transcripts from the s allele than from the S allele. Carriage of the s allele (Ss and ss genotypes) could be related to disturbances in the quantity of the COLIA1 mRNAs as in osteogenesis imperfecta (11). A change in the ratio of the produced collagen α1(I)/α2(I) chains was shown in osteoblasts from Ss heterozygotes (12). There is evidence that in osteogenesis imperfecta, an inactivating mutation in the COLIA1 gene results in production solely of α1(?)3 type I collagen (13). These data support the hypothesis that the s allele carriers have reduced bone strength. In this study, we examined the relation of the COLIA1 gene polymorphism with BMD in a random population-based sample of 400 Bulgarians. This is a pilot study examining the prevalence of this polymorphism and its association with forearm, lumbar spine and femoral neck BMD in a Bulgarian sample. We wanted to verify that this relation is the same as in other population groups.
Materials and Methods
The study included 400 women aged from 36 to 77 years old. Informed consent was obtained from each subject. All studied individuals were from Bulgarian ethnical origin (white Caucasians) and had no blood ties between them. About one third of the subjects were referrals to our Osteoporosis Centre by GPs or other practicing medical specialists. The second third were self-referrals and were measured because of their concern about possible osteoporosis based mainly on maternal history of low-trauma fractures and known risk factors. The remaining one third came from a population-based screening program of apparently healthy women living in an urban area. From all studied women, 220 who had low BMD (BMD T-score below -1.0) were labeled as cases and 180 were controls with normal BMD (T-score above -1.0). The rather high T-score threshold was chosen to separate the women with normal from those with abnormal BMD and, additionally, it increased the sensitivity of the genetic markers in identifying women with osteoporosis and fractures at the expense of specificity. All studied women were menopausal. They had no diseases causing secondary osteoporosis (endocrine, gastro-intestinal, liver diseases, kidney diseases, genetic or other etiologies). None of them received inhibitors of bone resorption, medications causing osteoporosis (glucocorticoids, immune suppressors, heparin, anticonvulsants, thyroid hormones and others), calcium or vitamin D. Physical activity was minimal in all participants. All fracture cases were confirmed by radiological reports at the Endocrinology Clinic and were part of patients’ disease characteristic. The study design was approved by the Ethics Committee at the Medical University Sofia as part of a research project. All patients gave their informed consent prior to inclusion in the study.
BMD was measured at the distal forearm by single-energy X-ray absorptiometry (SXA) on a DTX-100 Unit (Osteometer Meditech, USA) and at the lumbar spine (L1-L4) and femoral neck by dual-energy X-ray absorptiometry (DXA) on a Hologic QDR 4500 A device (Hologic Inc., Bedford, MA, USA). On the DTX-100, the distal region of interest begins at the 8-mm separation point between radius and ulna and then continues proximally for a distance of 24 mm (14). The ultra-distal site extends from the radial endplate proximally to the 8-mm point. BMD was measured according to the manufacturer’s instructions in g/cm2 separately for the distal (including radius and ulna) and the ultra-distal site (including only the radius). Z-scores were calculated automatically based on the manufacturer’s Danish database (issued 1994). BMD of the lumbar spine in the posterior-anterior (PA) projection was with a software version 8.26:3 (Hologic, Inc., Bedford, MA, USA). BMD L1-L4 was expressed in g/cm2 and, additionally, in terms of T- and Z-scores (15). The manufacturer’s American reference database was used (issued 1991). Standardization was performed daily by scanning a Hologic anthropomorphic phantom (for DXA) and a manufacturer-supplied forearm phantom (for SXA).
The polymorphic region is located on the long arm of chromosome 17 (17q22). DNA was isolated from whole blood. 100 ng of the DNA were used as template in the PCR reactions. dATP, dCTP, dTTP, dGTP - 1,25 mM each - were used to amplify intron 1 with Taq DNA polymerase. The PCR was performed in a DNA thermo cycler with a cycling protocol of 94ºC for 1 minute, 67ºC for 1.5 minute, 72ºC for 1 minute, for 35 cycles. The PCR product was then digested with MscI. Digestion of the product generated two fragments – 244 bp and 17 bp. Individuals homozygous for the SS genotype had a single uncut fragment of 261 bp, while homozygous for the ss genotype had two fragments (244 bp and 17 bp). The heterozygotes Ss had all three bands. PCR products were digested overnight with MscI and divided by electrophoresis on an acrylamide gel. Individuals were scored as SS, Ss, and ss according to the digestion pattern. Uppercase letters represent the absence and lowercase letters represent the presence of the restriction site.
All data were evaluated by the c2-test and presented as means±SD. The body mass index (BMI) was calculated as: BMI=weight (kg)/[height (cm)/100]2. The odds ratio (OR) for low BMD in the presence of a specific allele was calculated as: OR = (axd)/(bxc); where: a is the number of carriers among the cases, b is the number of non-carriers among the cases, c is the number of carriers among the controls, d is the number of non-carriers among the controls. The etiological factor (EF) showing what part of the condition (low BMD) might be attributable to the associated factor (Sp1 polymorphisms) on a population level was calculated as: EF = (OR-1)/OR=a/(a+b);
The polymorphism informative content (PIC) is an index of informative value of a genetic marker which takes into account the number of alleles and their frequencies. It is calculated according to the following formula: where pi is the frequency of the ith allele, and n is the number of alleles. The statistical significance of the differences was evaluated by c2-tests under the assumption of a normal distribution. The small study sample number did not allow either analysis for the adjusted effects of different factors on the COLIA1 genotype or multivariate analyses.
Genotype and Allele Frequencies
Among the studied 400 women (800 alleles), we found 408s alleles (51%). Among the 220 women with low BMD, 20 had the SS genotype, 94 – the Ss genotype, and 106 – the ss genotype. The alleles were as follows: 134 S and 306 s. Among the 180 controls with normal BMD, 93 had the SS genotype, 76 - the Ss, and 11 – the ss genotype. The alleles included 262 S and 98 s (Table 1). The observed allele frequencies in the controls (n=180) were: 73% for the S allele and 27% for the s allele. In low BMD subjects, the corresponding allele frequencies were 30% (for the S allele) and 70% (for the s allele). The differences were statistically significant after c2-test (p<0.05). The genotype frequencies in the healthy controls were: 52% - SS genotype, 42% - Ss genotype, and 6% - ss genotype. The corresponding genotype frequencies in cases with low BMD (n=220) were 9% (SS), 43% (Ss), and 48% (ss). If comparing genotype frequencies according to BMD, a higher prevalence of ss homozygotes (48%) was observed in cases with low BMD than among healthy controls (6%) together with a lower prevalence of SS homozygotes (9% in low BMD cases versus 52% in healthy controls). A higher prevalence of Ss heterozygotes was observed in low BMD cases (43%) compared with controls (42%). Therefore, the studied genetic marker is characterized by a high polymorphism information content (PIC=0.32) and by a high level of heterozygotes (Table 1).
Characteristics of the Study Subjects
The carriers of the three studied genotypes (SS, Ss, ss) did not differ in their age, years since menopause, current smoking and alcohol consumption (Table 2). A gene-dose effect on body weight was found. Body weight was lowest in the ss carriers and was medium in the Ss carriers (compared with the SS genotype). The mean BMI in the carriers of the SS genotype was higher than in the carriers of Ss (difference = 1.4 kg/m2, p=0.048). The mean BMI of the SS carriers was by 2.7 kg/m2 higher than BMI in ss carriers (p<0.01).
The Association Between Genotype and Previous Fractures
Table 2 shows the distribution of the reported osteoporotic fractures by the COLIA1 genotype. Fractures were twice more frequent among women with Ss genotype compared with SS genotype. Fractures were most frequent in women with ss genotype.
The Association Between Genotype and BMD
The three genotype groups differed in their BMD as measured at the forearm, lumbar spine and femoral neck (Table 2). At the three sites, BMD was highest in the SS carriers and was lowest in the ss group. Distal forearm BMD in ss carriers was 0.074 g/cm2 lower than in Ss carriers and 0.080 g/cm2 lower than in SS carriers (p<0.001). Ultradistal forearm BMD in the ss carriers was 0.084 g/cm2 lower than in Ss carriers and 0.101 g/cm2 lower than in SS carriers (p<0.01). The correlation between forearm BMD and genotype is shown in Figure 1. Lumbar spine BMD T-scores in ss carriers were by 0.5 standard deviations (corresponding to a BMD difference of 0.053 g/cm2) lower than those in Ss carriers and were by 1.0 standard deviation (BMD difference of 0.104 g/cm2) lower than those in SS carriers (p<0.001). The ss carriers had lower BMD and T-scores than Ss carriers (by 0.7 SD and 0.075 g/cm2, respectively) and than SS carriers (by 1.2 SD and 0.127 g/cm2) (p=0.01). The lumbar spine and femoral neck T-scores according to the three genotypes are shown in Figure 2.
Age-related Changes in the Association Between BMD and the Present Genotype
When dividing the study subjects into 10-year age groups, the differences in BMD between genotypes became small among younger women (from 40 to 50 y), but they grew bigger with increasing age among the three genotypes. E.g. in women 40-50 y of age, lumbar spine BMD in ss carriers was 3.4% (0.034 g/cm2) lower than in Ss carriers and 5% (0.050 g/cm2) lower than in SS (p=0.01). In women 51-60 y of age, these differences grew to 4% (0.039 g/cm2) and 8% (0.078 g/cm2), respectively (p=0.01). In women 61-70 y and older, the differences in lumbar spine BMD in the ss genotype compared to the Ss and SS genotypes reached 9% (0.091 g/cm2) and 16% (0.160 g/cm2), respectively (p<0.001). Therefore, our data show an increase in the gene-dose effect with age (Figure 3).
Genetic factors have an important role in determining BMD. Our study showed that the studied Bulgarian population differs significantly in the prevalence of the s allele from other European populations and is among the populations with the highest s allele frequencies. Since the s allele is associated with lower BMD, the populations where the allele frequency is lower are more “healthy” from an osteoporosis perspective. We compared the found allele frequencies with data for the general European population - 82% for the S allele and 18% for s allele, respectively (7). The s allele is found more often in Bulgarians than in other European populations. The genotype frequencies also differ from the published European data. In the study by Montanaro & Arciola (16), the SS, Ss and ss genotype frequencies were 67%, 30% and 3%, respectively. The observed differences are most significant when the ss group is concerned. We also showed an association of SS genotype with higher body weight (higher BMI). Uitterlinden et al. (7) reported that women with Ss genotype had lower body weight than those with SS genotype. In their study, the women in the ss group had the lowest body weight. These findings suggest that the association between COLIA1 genotype and BMD might be partly mediated by the genetic effect of COLIA1 on body weight. BMI in this case would have interfered as a confounding factor. The differences in the patients’ BMD values were however much greater than could be attributed solely to differences in BMI. We did not look for genotypes and BMD in the subgroups with equal BMI because of the relatively small study sample. The third relevant finding in our study is that the higher prevalence of osteoporotic fractures is associated with the s allele and the corresponding fracture risks were 52%, 22% and 10% in the ss, Ss and SS genotypes, respectively. A recent meta-analysis has confirmed the association of COLIA1-Sp1 polymorphisms with osteoporotic fractures in Caucasian postmenopausal women (17). The combined results showed that there was a significant difference in genotype distribution (SS odds ratio [OR] 0.72; Ss OR 1.18; ss OR 1.97) between patients with fractures and controls. Patients with vertebral fractures had a significantly higher frequency of Ss genotype and a lower frequency of SS genotype than controls; and those with nonvertebral fractures had a significantly higher frequency of ss genotype and a lower frequency of SS genotype than controls (17). Another study found that using the COLIA1 genetic marker could enhance the absolute fracture risk prognosis by improving risk classification by 2% for any fragility fracture, 4% for hip fracture, and by 5% for vertebral fracture, beyond age, BMD, prior fracture and propensity to falls (18). At last, we confirmed the relationship between COLIA1 genotype and BMD, as demonstrated in previous investigations in other populations (19,20). The Sp1 polymorphism in the COLIA1 gene was associated with osteoporosis at the lumbar spine in Mexican women (19). Uitterlinden et al. (7) reported BMD losses at the lumbar spine, femoral neck and the forearm (-25.6% / -25% / -18%) in SS carriers compared to -40% / -38% / -35% in ss carriers. A Danish study identified specific haplotypes associated with perimenopausal bone mass and loss, while other haplotypes were not (20). Genotype specific differences in BMD increased with age in the studied Bulgarian population group. Since the allele frequencies do not change with age, we may hypothesize that the studied polymorphism is most probably a marker of increased bone loss in elderly women and not a marker for lower peak bone mass. The odds ratio for low BMD and osteoporosis in the presence of COLIA1 risk marker (s allele carriers) is 10.69. The etiological factor of the marker is 0.82 showing that a big proportion of the disease on a population level is associated with the Sp1 polymorphism. Therefore, this polymorphism could be used as a marker for differences in bone quality and quantity. When considering our data, several study limitations should be kept in mind. First, our study population was of moderate size and quite heterogeneous which might be a source of bias. It was composed of patients referred by other physicians, self-referrals and women screened in a population-based program. The small sample number did not allow analysis for confounding variables such as BMI, years since menopause or multivariate analyses. Our study had insufficient power to analyze these populations separately. This heterogeneous population is, however, representative of the typical population receiving DXA scans at our hospital-based Osteoporosis Center. In addition, forearm BMD was measured at the distal and ultradistal sites, which are clearly different from the recommended by the International Society for Clinical Densitometry 33% radius site.
We found differences in forearm, lumbar spine and femoral neck BMD between homozygotes for the SS and ss COLIA1-Sp1 genotypes. The COLIA1 alleles were also associated with body weight. Studying the genotype by the COLIA1 polymorphism could provide additional information for the risk of osteoporotic fractures beyond the information obtained from bone densitometry. Further studies of larger cohorts and in ethnically diverse subgroups are still necessary.
No competing financial interests exist.
1. Kanis JA, Melton LJ, Christiansen C, Johnston CC, Khaltaev N. The diagnosis of osteoporosis. J Bone Miner Res 1994;9:1137 41.
2. Byers PH, Steiner RD. Osteogenesis imperfecta. Annu Rev Med 1992;43:269-82.
3. Grant SFA, Reid DM, Blake J, Herd R, Fogelman I, Ralston SH. Reduced bone density and osteoporosis associated with a polymorphic Sp1 binding site in the collagen type I α1 gene. Nat Genet 1996;14:203-5.
4. Bornstein P, McKay J, Morishima JK, Devarayalu S, Gelinas RE. Regulatory elements in the first intron contribute to transcriptional control of the human collagen α1(I) collagen gene. Proc Natl Acad Sci USA 1987;84:8869-73.
5. Langdahl BL, Ralston SH, Grant SFA, Eriksen EF. A Sp1 binding site polymorphism in the COLIA1 gene predicts osteoporotic fractures in both men and women. J Bone Miner Res 1998;13:1384-9.
6. Beavan S, Prentice A, Dibba B, Yan L, Cooper C, Ralston SH. Polymorphism of the collagen type I α1 gene and ethnic differences in hip-fracture rates. N Engl J Med 1998;339:351-2.
7. Uitterlinden AG, Burger H, Huang WJ, et al. Relation of alleles of the collagen type I α1 gene to bone density and risk of osteoporotic fractures in postmenopausal women. N Engl J Med 1998;338:1016-21.
8. Garnero P, Borel O, Grant SFA, Ralston SH, Delmas PD. Collagen I α1 Sp1 polymorphism, bone mass, and bone turnover in healthy French premenopausal women. The OFELY study. J Bone Miner Res 1998;13:813-7.
9. Hampson G, Evans C, Petitt RJ, et al. Bone mineral density, collagen type I α1 genotypes and bone turnover in premenopausal women with diabetes mellitus. Diabetologia 1998;41:1314-20.
10. Dean V, Smith FG, Robins SP, Ralston SH. Relationship between COLIA1 Sp1 alleles, gene transcription, collagen production and bone strength. Bone 1998;23:160.
11. Willing MC, Deschenes SP, Scott DA, et al. Osteogenesis imperfecta type I: molecular heterogeneity for COLIA1 null alleles of type I collagen. Am J Hum Genet 1994;55:638-47.
12. Mann V, Hobson E, Li B, Stewart T, et al. A COLIA1 Sp1 binding site polymorphism predisposes to osteoporotic fracture by affecting bone density and quality. J Clin Invest 2001;107:899-907.
13. Deak SB, van der Rest M, Prockop DJ. Altered helical structure of a homotrimer of α1(I) chains synthesized by fibroblasts from a variant of osteogenesis imperfecta. Coll Relat Res 1985;5:305-13.
14. Osteometer DTX-100. Product Documentation. Osteometer A/S, Rodovre, Denmark, 1994.
15. Hologic. Hologic QDR 4500 A - User’s guide. Hologic Inc. Waltham USA, 1996.
16. Montanaro L, Arciola CR. Detection of the G-T polymorphism at the Sp1 binding site of the collagen type Iα1 gene by a novel arms-PCR method. Genetic Testing 2002;6:53-57.
17. Ji G-R, YAO M, Sun C-Y, Zhang L, Han Z. Association of Collagen Type I α1 (COLIA1) Sp1 Polymorphism with Osteoporotic Fracture in Caucasian Post-menopausal Women: a Meta-analysis. J Int Med Res 2009;37:1725-32.
18. Tran BNH, Nguyen ND, Center JR, Eisman JA, Nguyen TV. Enhancement of Absolute Fracture Risk Prognosis with Genetic Marker: The Collagen I Alpha 1 Gene. Calcif Tissue Int 2011;85:379-388
19. Falcón-Ramírez E, Casas-Avila L, Miranda A, Diez P, Castro C, Rubio J, Gómez R, Valdés-Flores M. Sp1 polymorphism in collagen I α1 gene is associated with osteoporosis in lumbar spine of Mexican women. Miol Biol Rep 2011;38:2987-92.
20. González-Bofill N, Husted LB, Harsløf T, Tofteng CL, Abrahamsen B, Eiken P, Vestergaard P, Langdahl BL. Effects of COLIA1 polymorphisms and haplotypes on perimenopausal bone mass, postmenopausal bone loss and fracture risk. Osteoporos Int 2011;22:1145-56.