Meta-Analysis Identifies 13 New Genetic Regions Associated With Bone Mineral Density
By Jacquelyn K. Beals, PhD
Medscape Medical News
October 6, 2009 — Investigators have identified 20 genetic regions that influence bone mineral density (BMD) of the lumbar spine and femoral neck; 13 were not previously associated with BMD. The new study was a meta-analysis of 5 genomewide association studies (GWASs) in people of Northern European heritage. Of particular interest was the finding of skeletal site-specificity for 7 of the 20 regions.
The work, published online October 4 in Nature Genetics, was carried out by first author Fernando Rivadeneira, MD, PhD, from the Departments of Internal Medicine and Epidemiology, Erasmus Medical Center, Rotterdam, the Netherlands, and the Netherlands Consortium for Healthy Aging (NCHA), numerous colleagues in the Netherlands, and international collaborators.
Osteoporosis involves decreased bone mass along with bone deterioration, reducing bone strength and increasing the fracture risk. Typically, the incidence of osteoporosis rises with age. BMD serves as a clinical indicator for diagnosing osteoporosis and predicting the risk for fracture. Among skeletal sites where BMD is assessed are the lumbar spine and the femoral neck. This study combines data from 5 GWASs on BMD of these 2 sites into a meta-analysis of 19,195 subjects aged 18 to 96 years.
A threshold of alpha (probability of a type 1 error, or a false positive) = 5 × 10–8 was selected as the cutoff for GWAS significance. More than 400 single-nucleotide polymorphisms (SNPs) at 20 loci qualified for GWAS significance: 10 loci were associated with BMD of the femoral neck, 15 were associated with BMD of the lumbar spine, and 5 were associated with both sites. Thirteen of the 20 regions had never been associated with BMD at the GWAS level (4 were "suggestively associated" in previously published studies, but the remaining 9 were never reported in this context). The BMD associations of 7 regions were already recognized.
The following are examples of the newly reported loci:
A region on the short arm of chromosome 1 contains 2 SNPs highly associated with BMD of both the lumbar spine and the femoral neck (P from 2.63 × 10–13 to 1.8 × 10–12). Both SNPs lie within a gene involved in the "highly evolutionarily conserved Wnt signaling pathway" that influences differentiation and development of bone cells.
A region on the short arm of chromosome 7 contains a SNP strongly associated with BMD of lumbar spine (P = 1.1 × 10–9) but not significantly associated with BMD of the femoral neck (P = 8.9 × 10–4). A nearby gene influences endosomal cholesterol transport, and a study in Asian subjects found another SNP in this region associated with BMD in the heel, the tibia, and the radius.
The authors propose that "an underlying signal common to both Europeans and Asians might still be captured by these SNPs, considering the differences in [linkage disequilibrium] across populations."
A region on the short arm of chromosome 11 contains a SNP with strong site specificity, which is associated at the genomewide level with BMD of the femoral neck (P = 6.4 × 10–10) but not with BMD of the lumbar spine (P = .004). A nearby gene is highly expressed in skeletal muscle, and knockout mice lacking this gene activity demonstrate "mild skeletal abnormalities" related to the size of and mineral deposition in endochondral elements of the developing skeleton.
The study also examined the combined effect of the 20 loci on fracture risk. Together, the top 15 SNPs associated with BMD of the lumbar spine explained approximately 2.9% of the variance in this measurement; similarly, combining the top 10 SNPs associated with BMD of the femoral neck explains about 1.9% of the variance in that characteristic. The authors acknowledge that "these loci explain only a small fraction of the variance in BMD and hence an even smaller fraction of the heritability for fracture risk."
Nevertheless, knowledge of these variants illuminates pathways that contribute to variation in BMD. The independent effects observed for BMD of lumbar spine and femoral neck might reflect differences in the biological processes that contribute to the formation of cortical versus trabecular bone. "None of the SNPs . . . identified can be unequivocally designated as the underlying 'true' variants driving the associations," conclude the authors. However, they urge that the search for such variants should continue to enhance risk prediction and to translate into the development of new pharmaceuticals for the treatment of osteoporosis.
