Abstract 18806: Integrated Proteomic Analysis of Cardiovascular Disease Reveals Novel Protein Quantitative Trait Loci
Introduction: Recent genome-wide association studies (GWAS) have examined associations between genetic variants and gene expression to help understand the molecular underpinnings of complex diseases. Few large-scale studies have analyzed genetic variants that modify protein levels. Studying protein quantitative trait loci (pQTLs) may provide novel insights into the role of the proteome in the pathogenesis of disease.
Methods and Results: We assayed 71 plasma proteins in 7242 Framingham Heart Study participants (mean age 50 years, 53% women) using bead-based multiplex immunoassays. Proteins were selected using evidence of association with CVD or its risk factors based on: a) comprehensive literature search, b) proteomics discovery via mass spectrometry, and c) proteins coded by genes identified by gene expression profiling or published GWAS of CVD and its risk factors. We tested the associations of plasma concentrations of the 71 proteins with 25 million SNPs, adjusting for age, age2, sex, and family structure. We identified 18,549 genome-wide significant SNPs (P<5x10-8) for 57 proteins of which 36 had at least one cis-pQTL and 41 had at least one trans-pQTL. To explore how pQTLs may affect CVD phenotypes by altering protein levels, we examined the overlap of pQTLs with SNPs (P<5x10-8) from published GWAS results. Our analysis revealed that 210 pQTLs for 12 proteins coincided with SNPs for coronary artery disease (CAD) or myocardial infarction (MI). The top pQTLs for each protein often explained a substantial proportion of inter-individual variation in protein abundance. For example, rs12740374 at the CELSR2/SORT1 locus, reported in previous GWAS to be associated with lipids and CAD/MI, explained 15% of inter-individual variation in plasma granulin levels. Of note, progranulin binds to SORT1; Sort1 knockout mice have markedly elevated levels of progranulin. Granulin levels and CVD phenotypes share common genetic origins, suggesting that granulin may be a candidate for functional studies to develop novel ways to predict, prevent, or treat CVD.
Conclusions: We conducted GWAS for high-value plasma proteins, identified novel pQTLs, and elucidated integrated networks of genetic variants, circulating protein levels, and CVD traits.
- Cardiovascular disease
- Genome-wide association studies (GWAS)
- Coronary artery disease
Author Disclosures: G. Chen: None. C. Yao: None. S. Hwang: None. C. Liu: None. C. Song: None. T. Huan: None. M.M. Mendelson: None. P. Courchesne: None. A. Lyass: None. M. Larson: None. A.D. Johnson: None. D. Levy: None.
- © 2016 by American Heart Association, Inc.