Large-Scale Plasma Protein Profiling of Coronary, Carotid, and Peripheral Atherosclerosis

Atherosclerosis-related disease is the leading cause of death and disability worldwide. Atherosclerosis is considered a systemic disease, but symptoms and sequelae specific to the vascular bed also imply relevance of localized processes. This study aimed to identify plasma protein signatures associated with atherosclerosis in the coronary, carotid, and peripheral arteries, to reveal vascular bed-specific pathomechanisms and their relevance for the clinical outcome.

Data from the MyoVasc study, a prospective cohort study on chronic heart failure (NCT04064450) were analyzed. All individuals were profiled by plasma levels of 538 proteins using the proximity extension assay technology (Olink, Uppsala, Sweden). Elastic net-regularized logistic regressions, adjusted for age, sex, cardiovascular risk factors, and relevant medication intake, were used to select proteins associated with atherosclerosis in the three arterial beds, against a control group comprising atherosclerosis-free individuals as reference. A protein score (PS) was developed for each site based on the identified proteins. The PSs were subsequently analyzed using age- and sex-adjusted competing risk regressions for 3-point major adverse cardiovascular events MACE (incident myocardial infarction, stroke, cardiac death) with death from other than cardiac causes as competing event. Furthermore, the potential cellular origin of the identified plasma proteins was assessed using cell counts based on differential blood count (Hematology Analyzer, Sysmex, Kobe, Japan) and estimated cell fractions based on immune cell deconvolution of whole blood bulk RNA sequencing data (Illumina, San Diego, California, U.S.).

The analysis sample for the identification of protein signatures comprised 3,188 individuals (mean age: 64.6 ± 11.1; 35.9% women) with available proteomic data. This study identified 136 proteins that were associated with coronary atherosclerosis (case n=1,202; control n=822; leave-one-out cross-validated (CV) AUC: 0.958), 108 proteins related to carotid atherosclerosis (case n=1,683, CV-AUC: 0.868), and 127 proteins associated with peripheral atherosclerosis (case n=472, CV-AUC: 0.839). A subset of 34 proteins was detected in all three settings. The coronary (hazard ratio HR [95% CI]: 1.237 [1.106-1.384], p=0.0002), carotid (1.181 [1.056-1.320], p=0.0034), and peripheral PSs (1.276 [1.154-1.410], p<0.0001) all predicted 3-point MACE. Protein-protein interaction analysis and pathway enrichment analysis revealed shared and distinct pathomechanisms. Common processes, encompassing signaling by interleukins and extracellular matrix organization, were observed in all three atherosclerosis phenotypes. Triglyceride metabolism was more prominent in coronary atherosclerosis, whereas proteins involved in the coagulation cascade played a greater role in peripheral atherosclerosis. The carotid and coronary PSs were more strongly related to platelet activation (carotid PS: p<0.0001, coronary PS: p=0.0017), while the peripheral PS was more indicative of neutrophilic inflammation (p<0.0001) and regulatory T cell involvement (p=0.001).

This study identified similarities and differences in atherosclerosis-related plasma protein signatures between arterial beds, implying the involvement of shared and distinct pathomechanisms. The protein signatures were strong predictors of 3-point MACE, underscoring their clinical relevance.