Polymorphisms in the sclerosteosis/van Buchem disease gene (SOST) region are associated with bone-mineral density in elderly whites

Abstract

Osteoporosis has a strong genetic component, but the genes involved are poorly defined. We studied whether the sclerosteosis/van Buchem disease gene (SOST) is an osteoporosis-risk gene by examining its association with bone-mineral density (BMD). Mutations in SOST result in sclerosteosis, and alterations in the SOST gene expression may be causal in the closely related van Buchem disease. We used a set of eight polymorphisms from the SOST gene region to genotype 1,939 elderly men and women from a large population-based prospective-cohort study of Dutch whites. A 3-bp insertion (f=0.38) in the presumed SOST promoter region (SRP3) was associated with decreased BMD in women at the femoral neck (FN) (P=.05) and lumbar spine (LS) (P=.01), with evidence of an allele-dose effect in the oldest age group (P=.006). Similarly, a G variant (f=0.40) in the van Buchem deletion region (SRP9) was associated with increased BMD in men at the FN (P=.007) and LS (P=.02). In both cases, differences between extreme genotypes reached 0.2 SD. We observed no genotype effects on fracture risk, for the 234 osteoporotic fractures validated during 8.2 years of follow-up and for the 146 vertebral prevalent fractures analyzed. We did not find association between any of several frequent haplotypes across the SOST gene region and BMD. We did find evidence of additive effects of SRP3 with the COLIA1 Sp1 polymorphism but not with haplotypes of 3' polymorphisms in the vitamin-D receptor gene. The SOST-COLIA1 additive effect increased with age and reached 0.5 SD difference in BMD at LS in the oldest age group (P=.02). The molecular mechanism whereby these moderate SOST genotype effects are mediated remains to be elucidated, but it is likely to involve differences in regulation of SOST gene expression.

Figure 1

Structure of SOST gene region and overview of polymorphisms. The SRPs described in this study span ∼87.3 kb of genomic sequence. The two SOST exons are indicated by solid boxes, and the 52-kb VBD deletion region is indicated by the heavy horizontal dashed line. Nucleotide positions are relative to the SOST initiation codon (+1) and GenBank sequences AF326736 and AF397423. Polymorphisms identified in the course of this study are compared with those reported by Balemans et al. (2002a) and those present in the NCBI SNP database and are indicated below the gene map as well as by their presence (+) or absence (−). Lack of a plus sign (+) or minus sign (−) indicates that the corresponding region was not included in the study. The four most common haplotypes across SRP1–3 and SRP5–6 identified in the Rotterdam Study population are also shown, with their respective frequencies (where n=3,878 alleles).

Figure 2

Pairwise LD D′ values of eight SRPs across the SOST gene region. The software program PHASE was used to calculate values by use of true genomic distance (A) and by use of the SRPs at equal distance from each other (B).

Figure 3

LS BMD in women, stratified by tertiles of age and by SRP3 genotype. The three possible SRP3 genotypes are indicated by diamonds (DEL-DEL), squares (GGA-DEL), and triangles (GGA-GGA). P values for linear regression were adjusted for age, height, and weight. The breakdown of subjects aged 71–80 years was DEL-DEL, n=118; GGA-DEL, n=176; and GGA-GGA, n=51.

Figure 4

LS BMD in women, stratified by tertiles of age and by combined genotype for SRP3 and COLIA1 Sp1 polymorphisms. The four possible genotype groups are indicated by diamonds (reference = no risk allele), squares (COLIA1), triangles (SRP3), and circles (COLIA1 + SRP3). ANCOVA P values were adjusted for age, height, and weight. The breakdown of subjects aged 71–80 years was reference, n=76; COLIA1, n=41; SSRP3, n=147; and COLIA1 + SRP3, n=74.