Rare Coding Variants Are Associated with Osteoporotic Fracture: A Large-scale Exome-Chip Analysis of 44,130 Adult Caucasian Men and Women in CHARGE and GEFOS Consortia. Y. Hsu1,2,3, K. Estrada2,4, P. Leo5, A. Teumer6, C. Liu7, C. Medina-Gomez8, H. Zheng9, R. Minster10, LP. Lyytikäinen11, R. Pengelly12, R. Cruz Guerrero13, L. Oei8, M. Claussnitzer1, M. Kahonen14, C. Cooper15, A. Hannemann16, D. Karasik1, A. Uitterlinden8, LA. Cupples7, JA. Riancho Moral17, J. Holloway12, E. Duncan18, T. Lehtimäki11, T. Harris19, H. Wallaschofski16, B. Richards9, F. Rivadeneira8, M. Brown20, D. Chasman21, D. Kiel1 1) HSL Institute for Aging Research, Harvard Medical School, Boston, MA; 2) BROAD Institute of MIT and Harvard, Cambridge, MA; 3) Molecular and Integrative Physiological Sciences, Harvard School of Public Health, Boston, MA; 4) Analytic and Translational Genetics Unit, Massachusetts General Hospital, Boston, MA; 5) University of Queensland Diamantina Institute, Brisbane, Australia; 6) Interfaculty Institute for Genetics and Functional Genomics, University of Greifswald, Greifswald, Germany; 7) Biostatistics Dept. Boston University, Boston, MA; 8) Erasmus Medical Center, Rotterdam, Netherlands; 9) Departments of Medicine, Human Genetics, Epidemiology and Biostatistics, McGill University, Montreal, Canada; 10) Department of Human Genetics and Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA; 11) Department of Clinical Chemistry, Fimlab Laboratories, Tampere, Finland; 12) Human Genetics and Genomic Medicine, University of Southampton Faculty of Medicine, Southampton, UK; 13) University of Santiago de Compostela, Santiago de Compostela, Spain; 14) Department of Clinical Physiology, University of Tampere School of Medicine, Tampere, Finland; 15) University of Southampton, Southampton, UK; 16) Institute of Clinical Chemistry and Laboratory Medicine, Institute for Community Medicine, University Medicine Greifswald, University of Greifswald, Greifswald, Germany; 17) Hospital U.M. Valdecilla-IFIMAV, University of Cantabria, Santander, Spain; 18) Royal Brisbane and Women's Hospital, Brisbane, Queensland, Australia; 19) Intramural Research Program, National Institute on Aging, Bethesda, MD; 20) Diamantina Institute of Cancer, Immunology and Metabolic Medicine, Brisbane, Australia; 21) Brigham and Women's Hospital and Harvard Medical School, Boston, MA.

   Bone mineral density (BMD) is the most commonly studied skeletal phenotype, yet osteoporotic fracture (FX) is the most important clinical consequence of low BMD. Importantly, FX has a BMD-adjusted heritability of 35%~69% in Caucasians, suggesting its genetic architecture may not be the same as low BMD. Identifying associated and causal variants is a necessary step to study the underlying biology of FX risk. The availability of the exome genotyping chips with 236K coding variants (non-synonymous, splice sites and stop-altering SNPs selected from ~18,000 genes) provides a feasible way to identify potentially causal variants in the exome. We conducted an exome-chip meta-analysis to identify novel functional coding variants that are associated with FX risk. The exome-chip was genotyped in 10 cohort studies comprising 44,130 old adult participants (8,781 FX cases and 35,349 controls; 86% female). FX (excluding fingers, toes, skull, pathological fractures and those resulting from high trauma) phenotypes were obtained by interview and confirmed in most studies through either clinical or X-ray confirmation. Since observed MAF of most genotyped variants (78%) was 1%, we performed gene-based collapsing tests (allele-count and SKAT tests) in each study to identify rare coding variants (MAF 1%) associated with FX. Covariates adjusted in the models included age, age2, sex, weight, height, ancestral genetic background, site for multisite studies and estrogen use. For family-based studies, a kinship matrix was incorporated into test statistics. An inverse-variance fixed effect meta-analysis (seqMeta package) was used to combine results. The most significant association was found in the PPM1J gene (p=7.6x10-12). Other novel associations that achieved exome chip -wide significance (p < 4.2x10-6, Bonferroni correction) were found in WAC, DAZL, MRPS23 and SMPDL3B genes, which were not reported to be associated with BMD before. We also performed single variant association analysis for variants with MAF > 1%. The most significant, novel association was found for a missense SNP K450E in the SLFN14 gene (p=7x10-6). Other strong associations (p < 5x10-5) were found for previously reported missense variants V667M and A1330V in LRP5 and a common variant in SLC25A13. In summary, our analysis identified novel, rare, missense variants associated with FX. A large scale de-novo genotyping on selected variants in ~35,000 additional samples is underway to replicate these findings.

You may contact the first author (during and after the meeting) at