Searchable abstracts of presentations at key conferences on calcified tissues
Bone Abstracts (2017) 6 OC3 | DOI: 10.1530/boneabs.6.OC3

ICCBH2017 Oral Communications (1) (26 abstracts)

Principal component-derived bone density phenotypes and genetic regulation of the pediatric skeleton

Jonathan Mitchell 1, , Alessandra Chesi 1 , Shana McCormack 1, , Diana Cousminer 1, , Heidi Kalkwarf 3 , Joan Lappe 4 , Vicente Gilsanz 5 , Sharon Oberfield 6 , John Shepherd 7 , Andrea Kelly 1, , Babette Zemel 1, & Struan Grant 1,

1Children’s Hospital of Philadelphia, Philadelphia, PA, USA; 2University of Philadelphia, Philadelphia, PA, USA; 3Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, USA; 4Creighton University, Omaha, NE, USA; 5Children’s Hospital Los Angeles, Los Angles, CA, USA; 6Columbia University Medical Center, New York, NY, USA; 7University of California San Francisco, San Francisco, CA, USA.

Objectives: To determine if genetic variants associated with principal component-derived areal bone mineral density (aBMD) loading scores.

Methods: Our sample comprised 1,293 children of European ancestry enrolled in the longitudinal Bone Mineral Density in Childhood Study (52% female). The participants completed up to 7 annual study visits. From dual energy X-ray absorptiometry scans, sex and age-specific aBMD Z-scores were calculated for total hip, femoral neck, spine and distal radius. Principal components analysis, applied to the four Z-scores, generated new integrated aBMD phenotypes. Linear mixed effects models, adjusted for age, Tanner, BMI-Z, dietary calcium and physical activity, were used to test associations between a genetic score (percentage aBMD-lowering alleles carried at 63 GWAS-implicated loci) and the loading scores. We also performed a GWAS, using the baseline data, to identify loci associated with the loading scores.

Results: Four principal components (PC1-PC4) were identified that explained 68.1, 18.6, 10.5, and 2.8% of the variance, respectively. A higher PC1 loading score indicated higher bone Z-scores across all four sites. The genetic score was associated with lower PC1 loading score (beta=−0.05, P=3.9×10−10); from the GWAS, rs114260199 (LMO2/CAPRIN1, P=3.9×10−8) and rs75321045 (ZMAT4, P=2.5×10−8, females) were associated with PC1 loading score. A higher PC2 loading score indicated higher distal radius Z-score only. The genetic score was not associated with PC2; from the GWAS rs67991850 (CPED1, P=2.5×10−11) was associated with PC2 loading score. A higher PC3 loading score indicated higher spine Z-score only. The genetic score was not associated with PC3; from the GWAS rs58649746 (RAB11FIP5, P=4.8×10−9, females) was associated with PC3 loading score. A higher PC4 loading score indicated lower total hip Z-score, but higher femoral neck Z-score. No genetic associations were observed for PC4.

Conclusion: We identified four integrated pediatric aBMD phenotypes, including non-site-specific (PC1), distal radius-specific (PC2) and spine-specific phenotypes (PC3). An established genetic bone fragility score associated with the non-site-specific phenotype, but not the site-specific phenotypes. Novel variants near LMO2/CAPRIN1, ZMAT4, and RAB11FIP5 associated with non-site specific or spine specific phenotypes. These results highlight the utility of an integrated skeletal site phenotyping approach, which may help identify additional genetic loci associated with skeletal development.

Disclosure: The authors declared no competing interests.

Volume 6

8th International Conference on Children's Bone Health


Browse other volumes

Article tools

My recent searches

No recent searches.