A Proteomic Predictor of Lipoprotein Function and Coronary Artery Disease: From Conception to Launch Timothy Collier (Presenter) Quest Diagnostics - Cleveland HeartLab Presenter Bio: Dr. Timothy Collier is an Associate Scientific Director for Quest Diagnostics - Cleveland HeartLab Cardiometabolic Center of Excellence, where he oversees the technical performance of laboratory operations. Previously as a Sr. R&D Scientist, his responsibilities included the identification and development of assays for cardiovascular biomarkers. Prior to joining Cleveland HeartLab, he was a Post-Doctoral Research Scholar in Molecular Oncology and Hematology at Washington University School of Medicine in St. Louis, MO, focusing on the identification of drug-resistance mechanisms and structural studies of cancer-driving mutations using quantitative mass spectrometry.

Authors: Timothy S. Collier (1), Zhicheng Jin (1), Darlene L.Y. Dai (2), Virginia Chen (2), Zsuzsanna Hollander (2), Raymond T. Ng (2), Bruce M. McManus (2), Robert Balshaw (2), Sophia Apostolidou (3), Marc S. Penn (1), Cory Bystrom (1)
(1) Quest Diagnostics -Cleveland HeartLab Cardiometabolic Center of Excellence, Cleveland, OH (2) PROOF Centre of Excellence, Vancouver, BC, Canada (3) 3. Gynaecological Cancer Research Centre, Department of Women

High-density lipoproteins are complex, and cardiovascular health depends on their functional properties, not only their concentration. Cholesterol efflux capacity is one function that is associated with cardiovascular disease. However, the traditional cell-based assay lacks precision and scalability. Here we describe the conception, development, validation, and evaluation of a multi-marker proteomic method and accompanying algorithms for estimation of cholesterol efflux capacity (pCE) and prediction of coronary artery disease (pCAD). Measurement of both demonstrated consistent association with cardiovascular disease in multiple case-control studies. The precise and scalable nature of the method enables larger future studies to explore clinical utility.

Introduction

High-density lipoprotein (HDL) is a complex macromolecular complex that plays many roles in vascular biology. Cholesterol efflux capacity (CEC) is a measure of HDL function that has an inverse association with cardiovascular disease (CVD) (1-3). CEC is traditionally assessed by a complex, low-throughput, cell-based assay utilizing radioactive or fluorescent-labelled cholesterol to assess transport from macrophages into the extracellular medium (2,4). We hypothesized that CEC could alternatively be assessed based on the rapid and scalable quantitative assessment of the HDL-associated proteome and by extension, a novel predictor of cardiovascular disease might arise from such a metric. Here we describe the development, validation, and clinical evaluation of a multi-marker proteome panel for the assessment of CEC and prognosis of CVD risk.

Methods

We previously developed a means of rapidly enriching HDL-like lipoprotein particles from human serum by affinity to lipid-free, stable isotope labeled, His-tagged Apolipoprotein A-I (5). These ApoA-I associated lipoprotein particles (AALPs) were subjected to Lys-C digestion for subsequent LC-MS/MS analysis. A targeted multiple-reaction monitoring method for 21 HDL associated proteins was conducted on a set of 70 training and 35 test specimens using a 3-stream Agilent 1260 HPLC coupled to an Agilent 6495 triple quadrupole mass spectrometer using StreamSelect. Those proteins found to have a univariate correlation to cell-based CEC measurements (performed by Vascular Strategies, Inc.) were used to generate a multivariate OLS regression model for CEC (pCE). Using a nested case-control study of 157 CAD and 74 matched control specimens from the Fairbanks Institute biobank, logistic regression was utilized to re-weight coefficients in the pCE model to yield a new model predicting coronary artery disease (pCAD). The analytical method was then validated to CAP/CLIA standards and applied to another case-control study in 137 banked specimens from the United Kingdom Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) (6).

Results

Discovery using targeted LC-MS/MS analysis of the training and test samples yielded a pCE model composed of 5 proteins (Apolipoproteins A-I, C-I, C-II, C-III, C-IV; Spearman r=0.86). Evaluation of pCE in a case-control study of 231 specimens from healthy individuals and patients with CAD revealed lower pCE in CAD specimens (median: 9.91 %/4 hours) relative to controls (median: 10.20 %/4hours, P=0.03), consistent with previously published findings using the cell-based CEC measurement (1-2). The pCAD model was derived within this same case-control experiment to maximize cohort stratification based on diagnosis and demonstrated significantly improved classification (P<0.0001). Analytical validation demonstrated acceptable limits of quantification (LLOQ for individual proteins ranging from 0.2 nM to 15 nM), linear dynamic range, reagent recovery, imprecision (5.2%-14.0 % Lab CV), and stability. Following analytical validation of the multiplexed proteomic method, we conducted a case-control study of myocardial infarction in 137 UKCTOCS specimens. Results in the UKCTOCS study confirmed significant differences between cohorts; pCE was lower in cases (median=10.66, IQR 9.96 to 11.54) than in matched controls (11.15, IQR 10.48 to 12.32, P=0.015). The pCAD algorithm yielded lower risk scores for controls (median: 0.36, IQR 0.19 to 0.51) relative to case specimens (median: 0.51, IQR 0.38 to 0.64, P=0.001).

Conclusions & Discussion

We developed a multiplexed proteomic assay and demonstrated that it performs with satisfactory figures of merit. This work represents an incremental step toward demonstrating the feasibility of high-throughput multiprotein analysis in the clinical laboratory and presents further opportunities to explore the clinical validity and ultimately clinical utility of the pCE and pCAD biomarkers, including large cohort studies currently underway.

1. Khera, A.V. et al. N Engl J Med 2011; 364:127-135

2. Rohatgi, A. et al. N Engl J Med 2014; 317: 2383-2393

3. Saleheen, D. et al. Lancet Diabetes Endocrinol 2015; 3: 507-513

4. De la Llera-Moya, M. et al. Arteriosler Thromb Vasc Biol. 2010; 30: 796-801

5. Collier, T.S. et al. J Proteome Res 2018; 17: 1183-1193

6. Jin, Z et al. Clin Chem In-Press