ECTS Abstracts (2015) 1 P265

Bivariate Analyses of BMD and Lean Mass in Children Identifies Variants with Novel Pleiotropic Effects Across Six BMD Loci and in the TOM1L2 Locus

Carolina Medina-Gomez1,2, John P Kemp3,4, Denise H M Heppe2,5, Cindy G Boer1, Jon H Tobias6, Albert Hofman5, Vincent W V Jaddoe2,5, Andre G Uitterlinden1,2, David M Evans3,4 & Fernando Rivadeneira1,2

1Internal Medicine, Erasmus MC University, Rotterdam, The Netherlands; 2The Generation R Study Group, Erasmus Medical Center, Rotterdam, The Netherlands; 3University of Queensland Diamantina Institute, Translational Research Institute, Brisbane, Australia; 4MRC Integrative Epidemiology Unit, University of Bristol, Bristol, UK; 5Department of Epidemiology, Erasmus MC University, Rotterdam, The Netherlands; 6School of Clinical Sciences, University of Bristol, Bristol, UK.

Background: Lean and bone mass are heritable traits with high phenotypic correlation (rho=0.44), likely reflecting the underlying mechanical and biochemical interactions between tissues. Our aim was to estimate the shared SNP-heritability (genetic correlation) of both traits in children and identify genetic determinants displaying pleiotropic effects on lean mass and bone mass accrual.

Methods: Participants make part of two prospective population-based birth cohorts, the Generation R Study (GenR) and the Avon Longitudinal Study of Parents and Children (ALSPAC). GenR children (n=4,071) born in Rotterdam, Netherlands are of multiethnic background with mean age=6.2, SD=0.37 years. ALSPAC children (n=4,820) born in Avon, UK had mean age=9.9, SD=0.32 years. Lean mass and BMD were measured with DXA (GE-Lunar iDXA/ Prodigy) and genome-wide genotyping (Illumina 660K) imputed to HapMap. Shared heritability estimates derived from array data of GenR were obtained using GCTA (with modified admixed-aware relatedness estimates using REAP). GWAS in GenR and ALSPAC were run using PLINK multivariate. Meta-analysis was performed by Fisher’s method. All analyses were adjusted for age, sex, height, fat percent (and 20 genomic principal components in GenR). P<5x10−8 was considered genome-wide significant (GWS).

Results: SNP-heritability estimates were 0.31 for BMD and 0.40 for lean mass, with a genetic correlation of 0.3. The bivariate GWAS meta-analysis identified GWS associations with concordant effects on lean mass and BMD; mapping to six established BMD loci including: WNT4, GALNT3, CPED1/WNT16, RANKL, RIN3 and PPP6R3/LRP5. Another GWS signal mapping to the TOM1L2 locus, showed opposite loadings in lean mass (-0.46) and BMD (0.59). ENCODE analyses identified enhancers for SREBF1 in the same haplotype block.

Conclusion: Several variants at BMD loci exert pleiotropic effects on lean mass. Bivariate analysis is a powerful method for identifying novel pleiotropic effects. SREBF1 is a regulator of muscle protein synthesis down-regulating MYOD1, MYOG and MEF2C factors. Functional studies are required to unravel underlying pleiotropic mechanisms.

Disclosure: The authors declared no competing interests. This work was supported by the Netherlands Organization for Health Research and Development ZonMw VIDI 016.136.367.