Targeted Massively Parallel Sequencing of 128 Candidate Genes Reveals a Potential Molecular Cause in 72% of Cases with Idiopathic, Severe or Familial Osteoporosis

Presentation Number: SH01-5
Date of Presentation: April 1st, 2017

Manuela GM Rocha-Braz*1, Monica M Franca1, Adriana M Fernandes1, Regina M Martin1, Antonio M Lerario2, Berenice B Mendonca1 and Bruno Ferraz-de-Souza1
1Hospital das Clinicas, University of Sao Paulo School of Medicine, Sao Paulo, Brazil, 2Hospital das Clinicas da Faculdade de Medicina da Universidade de São Paulo, SAO PAULO

Abstract

Osteoporosis (OP) is a multifactorial disorder with high morbidity and mortality, and high heritability. Incorporating molecular information in clinical practice could lead to a better identification of individuals at risk for OP, decreasing disease burden. Several candidate genes for bone fragility have arisen, but the contribution of pathogenic variants in these genes to the pathophysiology of OP is still poorly understood. Our objective was to perform mutational analysis of candidate genes in a cohort of patients with idiopathic, severe or familial OP.

After IRB approval and informed consent, 29 cases (7 families and 22 sporadic) were selected, including men or premenopausal women with idiopathic OP or low bone density, and post-menopausal women with a T-score ≤-4.0 or a first-degree relative with idiopathic OP. All causes of secondary OP were excluded. Through a systematic review of monogenic diseases associated with bone phenotypes, reports of idiopathic or familial OP, and GWAS for OP or fracture, 128 candidate genes were identified and included in a massively parallel sequencing panel. Coding regions and 25-bp boundaries were captured with Agilent SureSelectXT and sequenced in Illumina NextSeq. Reads were aligned with BWA, and variant detection and copy-number variation analysis were performed with Platypus and CONTRA, respectively. Median coverage was 514x, and >98% targeted regions had a minimum coverage of 50x in all samples. Rare (MAF<0.005, ExAC & 1000Genomes) stop-gain, frameshift, splice site, or nonsynonymous variants predicted to alter protein function (SIFT, PolyPhen2, CADDphred) were considered potentially pathogenic variants (PPVs). All PPVs were confirmed by Sanger sequencing, and segregation analysis was performed when family control samples were available.

Thirty-five heterozygous PPVs were identified, 23% of which were in well-established candidate genes (COL1A1, COL1A2, WNT1, LRP5, PLS3), and 77% in new candidate genes (e.g. NBR1, AXIN1, WLS, LRP4). Only 1 PPV was found in LRP5, previously associated with several cases of idiopathic OP. In 8 cases (28%) no PPV was found. In 24% of the cohort, a combination of 2 or more PPVs were found, including a young woman with severe OP and PPVs on WNT1 and PLS3. PPVs were also found in GWAS candidate genes NBR1 and GPR68, associated with mineralization and osteoclast defects in mice, respectively, but without association with human OP phenotypes so far.

In conclusion, a targeted massively parallel sequencing strategy was able to identify potential molecular causes of OP in 72% of this cohort of extreme OP cases. Novel genes were associated to human OP, and digenic/oligogenic interactions were uncovered, expanding our knowledge of the molecular bases of bone strength. The identification of pathogenic variants associated with OP may allow personalized clinical management of patients and relatives in the future.

 

Nothing to Disclose: MGR, MMF, AMF, RMM, AML, BBM, BF