Proteomic signatures reveal why not all apolipoprotein B-containing lipoproteins are equal in atherosclerosis
J Morze, E Bjornson, M Y Mi, M Adiels, G E Melloni, N A Marston, M Guasch-Ferre, R E Gerszten, Q Sun, F B Hu, R Landberg, G Bergstrom, C J Packard, J Boren, C WittenbecherAbstract
Background/Introduction
All apolipoprotein B (apoB)-containing lipoproteins are causal risk factors for coronary artery disease (CAD). However, triglyceride-rich lipoproteins (TRL) and lipoprotein(a) [Lp(a)] exhibit greater per-particle atherogenicity than low-density lipoprotein (LDL) through mechanisms that remain incompletely understood.
Purpose
To identify plasma proteomic signatures specific to LDL, TRL, and Lp(a), and to determine whether these signatures mediate the differential contribution of each lipoprotein fraction to atherosclerotic risk.
Methods
First, we identified plasma proteins associated with LDL, TRL, and Lp(a) concentrations using multivariable-adjusted regression across 2918 proteins in UK Biobank (n=35,269). We further evaluated significantly associated proteins in two-sample Mendelian randomization to corroborate causality. Proteins consistently identified by both approaches were summarized into lipoprotein-specific multi-protein scores and tested for association with incident CAD (1599 events; median follow-up 13.7 years) using Cox regression. High-dimensional mediation analysis quantified the proportion of lipoprotein-associated CAD risk explained by proteomic alterations. Findings were replicated in the Multi-Ethnic Study of Atherosclerosis (n=5915).
Results
We identified 30 proteins associated with LDL, 471 with TRL, and 53 with Lp(a) in both observational and genetic analyses (Bonferroni-adjusted P<0.05). The TRL and Lp(a) signatures were largely distinct from LDL yet overlapped substantially with each other (36 shared proteins), with common enrichment in inflammatory pathways. Multi-protein scores for TRL (adjusted hazard ratio [HR] per SD 1.16; 95% CI 1.08–1.24) and Lp(a) (HR 1.09; 95% CI 1.03–1.15), but not LDL (HR 0.95; 95% CI 0.89–1.02), were associated with incident CAD independently of measured lipoprotein concentrations. Mediation analysis identified proteomic markers of immune activation and vascular remodeling as primary mediators of TRL-associated CAD risk (proportion mediated: 62%) and partial mediators of Lp(a)-associated risk (14%) (Figure 1).
Conclusions
Beyond arterial lipid deposition, TRL and Lp(a) converge on shared proteomic alterations in inflammatory signaling and vascular remodeling. These pathways represent potential therapeutic and monitoring targets to mitigate the excess cardiovascular risk conferred by these lipoprotein fractions.Figure 1.Mediation analysis results