Background: Whole-body bone mineral density (WB-BMD) is used to assess peak bone mass acquisition and bone health. It is highly heritable, with >50% of BMD variance accounted for by genetic factors. More than 80 genomic regions (loci) in the human genome have been associated with WB-BMD variation, and >500 with ultrasound-derived heel eBMD. However, underlying cell types and effector genes that regulate these loci remain largely unknown.
Aims: To conduct the largest genome-wide association study (GWAS) meta-analysis for WB-BMD and integrate the results with gene expression data from bone cells to identify genes and cellular mechanisms that regulate BMD.
Methods: GWAS meta-analysis included 11 cohorts and 107,000 individuals. WB-BMD associated loci were identified using GCTA-Cojo and deemed novel if located >1Mb from any known BMD-associated locus. Co-localisation analyses incorporating eQTL data from human osteoclast-like cells were used to screen WB-BMD loci and identify candidate effector genes. Effector genes were followed up in a single-cell RNA expression dataset encompassing 34 bone/marrow cell-types isolated from mice, as well the MGI mouse mutant database.
Results: 196 association signals for WB-BMD (P<5×10-8) were detected and 9 were novel. Co-localisation highlighted several potential effector genes acting through osteoclasts, including the V-ATPase proton pump subunit gene ATP6V1A (Fig.1A). The G-allele at the lead GWAS variant (rs2305545, MAF=0.35), was associated with increased WB-BMD and reduced ATP6V1A expression (Fig.1B). Differential gene expression analysis demonstrated up-regulation of ATP6V1A during differentiation of human osteoclast-like cells (P<0.0001, Fig.1C). Single-cell RNA-seq revealed that Atp6v1a is highly expressed in mouse osteoclasts relative to other cells (Fig.1D). Mice lacking Atp6v1a present with increased bone mineral content.
Conclusion: GWAS of WB-BMD identified 9 loci that have not been identified by much larger GWAS of ultrasound-derived heel eBMD. Co-localisation analysis implicated ATP6V1A as a potential effector gene that may act through osteoclasts to regulate BMD.