Proteomic Profiling of Early-Stage Heart Failure Identifies Candidate Biomarkers and Molecular Pathways of Cardiac Inflammation in the Project Baseline Health Study
Kottilil, K.; Nguyen, M.; Ramaker, M.; Cauwenberghs, N.; Kwee, L. C.; Short, S.; Plowman, S.; Mentz, R. J.; Douglas, P. S.; Holman, R. R.; Hernandez, A. F.; Mahaffey, K. W.; Haddad, F.; Shah, S. H.
Show abstract
BackgroundAsymptomatic structural heart disease, or stage B heart failure (HF), is clinically relevant for early HF diagnosis and prevention; circulating biomarkers could have prognostic significance for this. Thus, we performed proteomics discovery in a well-phenotyped cohort, the Project Baseline Health Study (PBHS). MethodsPBHS recruited participants with and without cardiovascular risk, collecting enrollment plasma biospecimens and cardiac imaging. Proteomic profiling (N=289) was performed using mass spectrometry on 503 individuals (185 stage B HF cases, 318 stage A HF controls). Logistic regression identified stage B-associated proteins, which were then eligible for inclusion in a joint protein score derived via elastic net. Scores were assessed for incident HF prediction in the UK Biobank (UKB) and the Exenatide Study of Cardiovascular Event Lowering (EXSCEL), examined for exenatide interactions in EXSCEL, and tested for imaging trait associations in PBHS and UKB. Mechanistic analyses of multivariate significant proteins included incident HF prediction in UKB and EXSCEL, Mendelian randomization (MR), and targeted methylation loci associations with stage B HF in PBHS. ResultsSixty-five proteins were associated with cases, of which 32 (49%) were retained via elastic net modeling; the resulting protein score showed good discrimination for cases versus controls (AUC 0.71, 95% CI [0.60, 0.82]). The protein score was significantly associated with incident HF in UKB and EXSCEL and left ventricular mass index (LVMI) in PBHS, as well as beneficially modified by exenatide in EXSCEL. Multivariate analysis prioritized 11 proteins associated with cases for mechanistic study, of which 4 (B2M, EFEMP1, CST3, HBB) showed significant incident HF associations in UKB and EXSCEL; CST3 and HBB were significant in MR. Methylation of cg08099136 in PSMB8 was significantly associated with cases in PBHS. ConclusionsOur findings highlight inflammatory biomarkers and mechanisms underpinning stage B HF. CpG hypomethylation at PSMB8 may be implicated in inflammation, resulting in increased LVMI (a known HF risk factor), with circulating B2M a byproduct of this process. ClinicalTrials.gov IdentifierNCT03154346
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Aptamer Proteomics for Biomarker Discovery in Heart Failure with Reduced Ejection Fraction 97%
- Genetically Predicted IL-18 Inhibition and Risk of Cardiovascular Events: A Mendelian Randomization Study 95%
- CD206 + IL-4Rα + MACROPHAGES ARE DRIVERS OF ADVERSE CARDIAC REMODELING IN ISCHEMIC CARDIOMYOPATHY 94%
Similar papers in this journal
Similar papers in this journal
- Deciphering human heart failure with preserved ejection fraction (HFpEF) at single cell resolution 95%
- Genome wide association analysis in dilated cardiomyopathy reveals two new key players in systolic heart failure on chromosome 3p25.1 and 22q11.23 94%
- Hypoferremia reduces long-term risk of major adverse cardiovascular events after STEMI by averting the myocardial reactive iron storm 92%
Similar papers in this journal
- Development and validation of imaging-free myocardial fibrosis prediction models, association with outcomes, and sample size estimation for phase 3 trials 95%
- High Variability of Body Mass Index Independently Associated with Incident Heart Failure 94%
- Loss of cardiac PFKFB2 drives Metabolic, Functional, and Electrophysiological Remodeling in the Heart 93%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.