Aptamer-Based Proteomic Platform Identifies Novel Protein Predictors of Incident Heart Failure and Echocardiographic Traits
Nayor, Matthew; Short, Meghan I.; Rasheed, Humaira; Lin, Honghuang; Jonasson, Christian; Yang, Qiong; Hveem, Kristian; Felix, Janine F.; Morrison, Alanna C.; Wild, Philipp S.; Morley, Michael P.; Cappola, Thomas P.; Benson, Mark D.; Ngo, Debby; Sinha, Sumita; Keyes, Michelle J.; Shen, Dongxiao; Wang, Thomas J.; Larson, Martin G.; Brumpton, Ben Michael; Gerszten, Robert E.; Omland, Torbjørn; Vasan, Ramachandran S.
Peer reviewed, Journal article
Published version
View/ Open
Date
2020Metadata
Show full item recordCollections
Original version
10.1161/CIRCHEARTFAILURE.119.006749Abstract
Abstract
Background:
We used a large-scale, high-throughput DNA aptamer-based discovery proteomic platform to identify circulating biomarkers of cardiac remodeling and incident heart failure (HF) in community-dwelling individuals.
Methods:
We evaluated 1895 FHS (Framingham Heart Study) participants (age 55±10 years, 54% women) who underwent proteomic profiling and echocardiography. Plasma levels of 1305 proteins were related to echocardiographic traits and to incident HF using multivariable regression. Statistically significant protein-HF associations were replicated in the HUNT (Nord-Trøndelag Health) study (n=2497, age 63±10 years, 43% women), and results were meta-analyzed. Genetic variants associated with circulating protein levels (pQTLs) were related to echocardiographic traits in the EchoGen (n=30 201) and to incident HF in the CHARGE (n=20 926) consortia.
Results:
Seventeen proteins associated with echocardiographic traits in cross-sectional analyses (false discovery rate <0.10), and 8 of these proteins had pQTLs associated with echocardiographic traits in EchoGen (P<0.0007). In Cox models adjusted for clinical risk factors, 29 proteins demonstrated associations with incident HF in FHS (174 HF events, mean follow-up 19 [limits, 0.2–23.7] years). In meta-analyses of FHS and HUNT, 6 of these proteins were associated with incident HF (P<3.8×10−5; 3 with higher risk: NT-proBNP [N-terminal proB-type natriuretic peptide], TSP2 [thrombospondin-2], MBL [mannose-binding lectin]; and 3 with lower risk: ErbB1 [epidermal growth factor receptor], GDF-11/8 [growth differentiation factor-11/8], and RGMC [hemojuvelin]). For 5 of the 6 proteins, pQTLs were associated with echocardiographic traits (P<0.0006) in EchoGen, and for RGMC, a protein quantitative trait loci was associated with incident HF (P=0.001).
Conclusions:
A large-scale proteomics approach identified new predictors of cardiac remodeling and incident HF. Future studies are warranted to elucidate how biological pathways represented by these proteins may mediate cardiac remodeling and HF risk and to assess if these proteins can improve HF risk prediction.