Proceedings of the 9^th HDL Workshop: Focus on Cardiovascular Disease May 16-17, 2019 Boston, MA

Abstract

The HDL Workshop was established in 2009 as a forum for candid discussions amongst academic basic scientists, clinical investigators, and industry researchers about the role of HDL in cardiovascular disease. This 9th HDL Workshop was held on May 16–17, 2019 in Boston, MA, and included outstanding oral presentations from established and emerging investigators. The Workshop featured five sessions with topics that tackled the role of HDL in the vasculature, its structural complexity, its role in health and disease states, and its interaction with the intestinal microbiome. The highlight of the program was awarding the Jack Oram Award to the distinguished professor emeritus Godfrey S. Getz from the University of Chicago. The 10th HDL Workshop will be held in May 2020 in Chicago, and will continue the focus on intellectually stimulating presentations by established and emerging investigators on novel roles of HDL in cardiovascular and noncardiovascular health and disease states.

Introduction

The HDL Workshop was established in 2009 to advance our understanding of high-density lipoprotein (HDL) biology (1). In clinical medicine, the measurement of the cholesterol (C) component of HDL, HDL-C, remains a standard-of-care requirement for cardiovascular disease (CVD) risk prediction. In 2019, the American Heart Association (AHA) and other professional societies issued an updated guideline on cholesterol management for primary and secondary CVD prevention (2). This document targeted clinical management of LDL-C and non-HDL-C, while providing no updates on HDL, HDL-C, or HDL functional metrics in CVD risk assessment. The AHA and the American College of Cardiology (ACC) did provide an online risk calculator to help clinicians assess CVD risk in their patients (3). The risk factors included in the calculator are gender, age, race, total cholesterol, LDL-C, HDL-C, statin treatment, systolic blood pressure, hypertension treatment, history of diabetes, current smoker, and aspirin therapy. Another well-known CVD risk calculator, the Framingham Risk Score, also incorporates HDL-C as a component in the 10-year risk assessment and assigns a score of −1 if patients have HDL-C ≥ 60 mg/dl (4).

The latter would suggest that higher HDL-C levels are cardioprotective but this paradigm is now being challenged, and many argue for a reassessment of HDL-C in these risk calculators. Many clinical investigators agree that the results of recent genetic association studies as well as large clinical trials targeting increases in HDL-C to reduce CVD risk as a primary endpoint have been disappointing (5–8). Such observations leave the field wide open in better understanding the complexities of HDL biology and where this research might be of help to clinicians. A major objective of the HDL Workshop has been, and will continue to be, a forum for academic basic scientists, clinicians, and industry researchers to share their research findings and perspectives. Over the past ten years attendance has steadily grown demonstrating that interest in HDL biology is alive and well. Therefore, we now present the first of many future HDL Workshop Proceedings with the hopes of engaging scientists and clinicians to attend and present their research on this fascinating topic.

In this year’s program, the HDL Workshop took place in Boston, MA over a two-day period. It included oral presentations from invited established investigators as well as short talks from trainees. A new feature was a poster presentation session. The highlight of the event was the annual Jack Oram Research Award, honoring the late Dr. Jack Oram and generously funded by the Dyslipidemia Foundation of Boston. This year’s award was presented to the distinguished professor emeritus Dr. Godfrey S. Getz from the University of Chicago.

HDL Workshop Program

The program was chaired by Dr. Catherine Martel and co-chaired by Dr. Annabelle Rodriguez. Over the course of the two-day program there were five sessions of oral presentations. The topics of each session were the following: HDL and the Vasculature, HDL Structural Complexity, Clinical Perspectives on HDL Measurement, HDL in Health and Disease, and Microbiota and HDL Metabolism. Please note that each of the following abstracts were written by the presenter with references included for the reader’s benefit.

Session One: HDL and the Vasculature

1. Bernardo L. Trigatti, Ph.D.

Professor, Department of Biochemistry and Biomedical Sciences, McMaster University and Thrombosis and Atherosclerosis Research Institute, McMaster University and Hamilton Health Sciences, Hamilton, Ontario, Canada. HDL and SR-B1 in Cardiovascular Disease: Atherosclerosis and Beyond.

Global knockout of the HDL receptor, scavenger receptor class B type I (Scarb1 gene, SR-B1 protein), in atherogenic mice increases their susceptibility to high fat diet induced aortic sinus atherosclerosis (9). In addition, when fed highly atherogenic, high fat and/or high cholesterol diets, Scarb1/Ldlr double knockout mice develop occlusive coronary artery atherosclerosis, extensive myocardial fibrosis and early death (10). The extensive myocardial fibrosis is undoubtedly triggered by the occlusive coronary atherosclerosis in these mice, but we wondered if it might also reflect a role for SR-B1 in cardiomyocytes themselves. SR-B1 is expressed in cardiomyocytes, and SR-B1’s ligand, HDL, is able to protect cultured mouse and human cardiomyocytes from apoptosis induced by the anti-cancer drug doxorubicin (11). This protection is lost when SR-B1 is inactivated. HDL induces phosphorylation of Akt in mouse and human cardiomyocytes in an SR-B1 dependent manner and Akt1 or Akt2 knockdown impair HDL-mediated protection against doxorubicin mediated apoptosis (12). In vivo, treatment of mice with doxorubicin over five weeks triggers greater cardiomyocyte apoptosis in Scarb1^−/− than wild type (WT) mice. Transplantation studies implicate SR-B1 in the heart itself. Finally, human apolipoprotein A-I (apoA-I) transgenic overexpression or injection protect WT but not Scarb1^−/− mice against doxorubicin induced cardiomyocyte apoptosis and cardiac dysfunction (11–12). In conclusion, while global knockout studies clearly demonstrate a protective role for SR-B1 against atherosclerosis (likely by virtue of its function in liver mediated reverse cholesterol transport), SR-B1 and HDL appear to have cardioprotective roles beyond atherosclerosis. Notably, this includes protection of cardiomyocytes against cytotoxicity by the chemotherapeutic agent doxorubicin.

2. Chieko Mineo, Ph.D.

Associate Professor, Center for Pulmonary and Vascular Biology, Department of Pediatrics and Cell Biology, University of Texas Southwestern Medical Center, Dallas, Texas. SR-B1 and Atherosclerosis: A New Paradigm for an Old Story.

The biology of the high affinity HDL receptor SR-B1 is best understood in the liver, where the receptor plays a key role in reverse cholesterol transport, facilitating cholesterol transfer from HDL to hepatocytes for disposal in bile (13). The global SR-B1 knockout mice show exaggerated atherosclerosis in the setting of hypercholesterolemia. In addition to the liver, SR-B1 is expressed in multiple cell types that are relevant to atherogenesis, including macrophages, platelets and endothelium. In endothelial cells, SR-B1 has been shown to promote production of the anti-atherogenic molecule nitric oxide (14). To study how endothelial SR-B1 impacts atherosclerosis, we generated mice lacking SR-B1 selectively in the endothelium by crossing floxed Scarb1 mice with VE Cadherin Cre mice (Scarb1^ΔEC) (15). Contrary to our prediction, Scarb1^ΔEC mice were protected from hyperlipidemia-induced atherosclerosis without affecting the plasma lipid profiles. Recognizing that SR-B1 binds both HDL and LDL, and that LDL uptake into the vascular wall is the initial step of atherogenesis, we hypothesized that SR-B1 facilitates LDL transport through the endothelial layer. Using multiple loss-of-function approaches in vivo, we demonstrated that SR-B1 promotes LDL uptake into the aorta. Studies in cultured endothelial cells further revealed that transendothelial transport of LDL requires SR-B1 C-terminal cytoplasmic domain that recruits DOCK4. DOCK4 initiates LDL-SR-B1 internalization via activation of the small GTPase Rac1. Our study revealed that expression of SR-B1 and DOCK4 is elevated in human atherosclerotic arteries compared with normal arteries. These findings suggest that the mechanisms that influence LDL transport across endothelium might be targeted to provide novel interventions to prevent atherosclerosis.

3. Catherine Martel, Ph.D.

Associate Professor, Canada Research Chair in Lymphatics and Cardiovascular Medicine, Department of Medicine, Université de Montréal, Montreal Heart Institute, Montreal, Quebec, Canada. Extracellular Vesicles at the Heart of Lymphatic Function.

Lymphatic vessels (LVs) form a whole body network that maintains fluid balance, dietary lipid absorption and pathogen clearance from peripheral tissues. In the past years, the direct role of LVs in atherosclerosis has been more fully assessed. We reported that functional adventitial LV are essential to first mobilize cholesterol out of the vessel wall before reaching the circulation (16). A defect in the propelling capacity of the collecting LV, rather than a defect in the absorptive capacity of the lymphatic capillaries would be the instigating element of this atherosclerosis-related lymphatic dysfunction (17). We are now pursuing the mechanisms that contribute to the loss of collecting LV function during atherosclerosis, and how can they be modulated. Cell fragments called extracellular vesicles (EVs) are important cell-cell messengers released upon cell activation or death and are found in atherosclerotic lesions. We reported for the first time that EVs of heterogeneous origins are also present in lymph, and that these small vesicles are even more abundant in atherosclerotic mice (18). With preliminary data suggesting that specific subsets of EVs could be harmful to the lymphatic endothelium (19), we hypothesize that EVs accumulation could be, at least in part, responsible for the atherosclerosis-related lymphatic dysfunction. This massive accumulation of EVs in the artery wall could in turn be due to the underlying reduced LV function. Altogether, these data suggest that EVs might be the connection between abnormal lymphatic function and atherosclerosis progression.

4. Darcy Knaack, Ph.D. graduate student

(Dr. Daisy Sahoo, mentor), Department of Biochemistry, Division of Endocrinology, Medical College of Wisconsin, Milwaukee, Wisconsin. The roles of SR-B1 and CD36 in Maintenance of Macrophage Cholesterol Homeostasis.

Atherosclerosis is a chronic inflammatory disease characterized by buildup of cholesterol-rich macrophages within the endothelium. Two scavenger receptors that regulate macrophage cholesterol homeostasis and cholesterol transport are SR-B1 and CD36, which bind HDL and oxidized LDL (oxLDL), respectively. Their individual roles in modulating atherosclerosis have been widely studied, but how they influence each other’s functions has yet to be investigated (20–21). To test our hypothesis that SR-B1 and CD36 are functional partners that mediate macrophage cholesterol homeostasis, we performed four different assays in peritoneal macrophages from WT, Scarb1^−/−, and Cd36^−/− mice. First, co-immunoprecipitation assays demonstrated a potential interaction between SR-B1 and CD36 that was enhanced with oxLDL treatment in WT macrophages. Next, we examined how cholesterol transport functions of these receptors were affected when one receptor was absent. We provide evidence that when CD36 is absent, the ability of HDL to bind surface receptors was impaired, as was the ability of the cell ability to internalize HDL-cholesterol as compared to WT or Scarb1^−/− mice. We further show that membrane distribution of cholesterol is altered when either receptor is absent. Lastly, we used sucrose-gradient fractionation to examine whether absence of SR-B1 or CD36 impacts membrane localization of the other receptor. Our data suggest that, in WT cells, SR-B1 and CD36 reside within lipid raft microdomains, but when one receptor is absent, the other migrates to non-raft domains. Based on these observations, there appears to be a cooperative partnership between SR-B1 and CD36, suggesting they may work together to promote macrophage cholesterol homeostasis.

Session Two: HDL Structural and the Vasculature

1. John T. Wilkins, M.D.

Associate Professor, Department of Medicine, Division of Cardiology, Northwestern University, Chicago, Illinois. Henrique Seckler, M.S. Departments of Chemistry and Molecular Biosciences, the Chemistry of Life Processes Institute, and the Proteomics Center of Excellence, Northwestern University, Evanston, Illinois. Differential ApoA-I Proteoform Quantification across HDL Particle Size Subtypes.

HDL particle size subtypes differ in proteome, lipidome, and functional characteristics (22). We recently reported that apoA-I has specific post-translational modifications that occur at precise amino acid residues, yielding 14 distinct specific molecular forms (proteoforms) of apoA-I (23). Fatty-acid-modified proteoforms of ApoA-I (Acyl-K88 apoA-I) are associated with HDL efflux, but it is unknown if apoA-I proteoforms vary with HDL particle size. We hypothesized that HDL subtypes have distinct apoA-I proteoform profiles. Using pooled samples from 30 healthy Coronary Artery Risk Development in Young Adults (CARDIA) participants we adapted CN-GELFrEE, a native acrylamide-gel electrophoretic technique, to separate HDL particles by size. We used western blot to quantify apoA-I content and average particle size of the electrophoretic fractions and then submitted apoA-I-containing fractions to liquid-chromatography mass spectrometry for proteoform quantification. We observed 36 distinct HDL size fractions covering the size range of HDL (5–11nm). Proteoform quantification revealed a significantly higher proportion of Acyl-K88 apoA-I in the pre-beta-1 (5–7.5nm) and alpha-1 (9–11nm) size ranges in comparison to medium size subtypes (7.5–9nm) [mean fold difference (MFD): 1.4x and 1.2x, respectively]. Glycated and oxidized proteoforms were significantly more abundant (MFD: 1.7x and 1.8x, respectively) in the medium size ranges. We concluded that CN-GELFrEE is a novel, high-resolution HDL size separation technique and is compatible with downstream proteomics. These data suggest the profile of apoA-I proteoforms in HDL particles is size-related and that apoA-I proteoforms may be important markers or mediators of different functional characteristics across HDL size subspecies.

2. Daisy Sahoo, Ph.D.

Professor of Medicine, Division of Endocrinology, Medical College of Wisconsin, Milwaukee, Wisconsin. The importance of HDL clearance: lessons from Dr. Jekyll and Mr. Hyde.

HDL protects against CVD, primarily due to its role in delivering peripheral cholesterol to the liver for excretion via reverse cholesterol transport (RCT). The last step in RCT requires the binding of HDL to SR-B1. Paradoxically, despite high circulating levels of HDL-C, humans harboring common and rare SR-B1 (SCARB1 gene) mutations have higher CVD risk, paralleling the failure of HDL-raising therapeutics to protect against CVD (5–8, 24–25). Thus, the field is appreciating that efficient cholesterol flux via RCT is critical to maintaining the protective functions of HDL. We tested the hypothesis that impaired HDL clearance promotes HDL dysfunction through increased oxidative modification. We observed that HDL modified by reactive aldehydes such as acrolein and malondialdehyde (MDA) impaired the athero-protective functions of HDL in macrophages. Compared to native HDL, acrolein- and MDA-modified HDL have impaired abilities to promote migration of primary murine peritoneal macrophages. Further, MDA-HDL increased the ability of macrophages to generate reactive oxygen species. Given these findings, our laboratory took a structure-function approach to understand how the SR-B1-HDL interaction can enhance cholesterol clearance via RCT. In our recently solved NMR structure of an SR-B1 peptide, we have identified a short amphipathic alpha-helix that, when mutated along its hydrophobic face in full-length SR-B1, exhibits impaired HDL binding and delivery of HDL-C into cells. Using electron paramagnetic resonance spectroscopy, we are testing whether this region is a juxtamembrane helix that mediates membrane/receptor interactions to facilitate cholesterol transport. Together, an improved understanding of SR-B1 structure will help us develop strategies to enhance RCT and reduce CVD risk.

3. Bela Asztalos, Ph.D.

Scientist I, Human Nutrition Research Center,Tufts University, Boston, Massachusetts. Analyses on HDL composition and function.

The composition and functional relationship of HDL particles is poorly understood. We tested the hypothesis that the functionality of HDL particles is significantly influenced by their lipid composition. We determined apoA-I and eight lipid classes in five different-size HDL subpopulations isolated from coronary heart disease (CHD) patients having low preβ−1 functionality (as defined by preβ−1-concentration-normalized cholesterol-efflux capacity) and control subjects having either low or high preβ−1 functionality. The average number of apoA-I and lipid molecules were 4 and 426 in α−1, 3 and 227 in α−2, 2 and 112 in α−3, 2 and 96 in α−4, and 2 and 57 in preβ−1, respectively. Interestingly, one third of the lipid molecules in the discoid-shaped preβ−1 particles were “core” lipids (CE and TG). The phospholipid/CE and free-cholesterol/CE ratios were significantly higher in preβ−1 compared to α-mobility particles. The preβ−1 phospholipid/CE ratio was significantly higher in low preβ−1 functionality subjects independent of the presence or absence of CHD. There were strong positive correlations between preβ−1 particle functionality and the abundance of major lipids in subjects with high preβ−1-functionality but not in subjects with low-preβ−1-functionality and/or CHD. The lipid composition of the large-HDL particles was significantly different between the CHD and the control groups. In CHD patients, the high TG/CE and phospholipid/CE ratios were associated with increased functionality of large-HDL particles. These data do not support clear causative relationships between the lipid composition and the functionality of HDL particles. Considering these results we hypothesize that other factors, for example the lipid-binding capacity of apoA-I, regulate HDL particle functionality.

Session Three: Clinical Perspectives on HDL measurement

1. Annabelle Rodriguez, M.D.

Professor, Cell Biology, Linda and David Roth Chair of Cardiovascular Research, Center for Vascular Biology, University of Connecticut Health, Farmington, Connecticut. High HDL-C Paradox: A Clinical Vignette.

A clinical vignette was presented to highlight the case of a 52 year old woman with elevated HDL-C (110 mg/dl) and elevated LDL-C (170 mg/dl) levels with a positive family history of premature CVD. The patient had no known personal history of CVD, smoking or history of diabetes mellitus, but on physical examination was found to have bilateral carotid bruits. The patient was reluctant to start statin therapy as she and her primary care physician believed that the HDL-C levels would provide cardioprotection. This case demonstrated the high HDL-C paradox and the need for more research in this group of patients to uncover the underlying genetic and non-genetic etiologies of increased CVD risk (24, 25).

2. Samia Mora, M.D.

Director, Center for Lipid Metabolomics, Associate Professor of Medicine, Department of Medicine, Divisions of Preventive and Cardiovascular Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts. Clinical perspectives on HDL measurement.

While HDL-C is a strong inverse indicator of CVD risk, this has not yet translated into clinical benefit. Given the extreme heterogeneity of HDL structure and function, measuring only the cholesterol content of HDL will, at best, only partially reflect the potential role of HDL in CVD risk assessment and therapeutic drug development. This has led to interest in developing HDL metrics that might better indicate the atheroprotective functions of HDL. Proposed measurements include HDL particle number (HDL-p), average size, subclasses, and functional assays. Of the alternate HDL metrics, the number of HDL particles (HDL-p) has potential utility in CVD prevention. At present it can be measured clinically by NMR spectroscopy (LabCorp) and ion mobility analysis (Quest Diagnostics). Most studies have been based on the NMR method. Data from the Multi-Ethnic Study of Atherosclerosis and the placebo arm of the Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) showed that HDL-C is no longer predictive of CVD after adjusting for HDL-p, while HDL-p remained inversely associated with CVD after adjusting for HDL-C (26–27). HDL-p might also be of potential value in targeting residual CVD particularly for primary prevention as indicated by data from the rosuvastatin arm of JUPITER, where HDL-p, but not HDL-C, was inversely predictive of CVD (28).

2. Marina Cuchel, M.D., Ph.D

Research Associate Professor, Department of Medicine, Division of Translational Medicine and Human Genetics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania. Clinical perspectives on HDL measurement.

Low levels of HDL-C are associated with increased CVD risk. However, contrary to LDL-C, the HDL-C association with CVD risk is not linear and, in recent years, several data have emerged highlighting the absence of protection against CVD, or even an increase in CVD risk, at high HDL-C concentrations, suggesting the presence of a U-shaped curve between HDL-C and all-cause mortality (29–30). Furthermore, HDL-C is not an adequate biomarker of HDL function and several additional biomarkers are currently being investigated. Reverse cholesterol transport (RCT) is one of the most important athero-protective functions of HDL. Cholesterol efflux capacity, an in vitro assay assessing the ability of HDL to efflux cholesterol from macrophages (the first step of RCT), has been inversely associated with both prevalence and incidence of CVD in several studies (31–32). Unfortunately, it is not available in routine clinical care and its results may be affected by underlying conditions such as diabetes and an inflammatory status. Recently, a method to assess macrophage specific reverse cholesterol transport in vivo has been developed that could be of use during drug development (33). Excitingly, thanks to the greater awareness of its remarkable heterogenicity in size and composition, our understanding of HDL’s role in the physiopathology of many diseases and conditions is expanding beyond its role in the progression of atherosclerosis and RCT.

Session Four: HDL in Health and Disease

1. Henry J. Pownall Ph.D.

Professor of Bioenergetics, and Corina Rosales Ph.D., Assistant Professor of Bioenergetics, Institute for Academic Medicine, Houston Methodist, Weill Cornell Medical College, Houston, Texas. HDL and serum opacity factor.

The optimal range for a low risk of atherosclerotic cardiovascular disease (ASCVD) among humans is HDL-C between 60–80 mg/dL with higher and lower concentrations associated with higher risk. Whereas low plasma HDL-C levels are associated with metabolic syndrome, the cause of ASCVD among patients with high plasma HDL-C levels is unknown and interventions that rescue HDL function have not been identified. Scarb1^−/− mice, which have high plasma HDL levels present with multiple metabolic abnormalities; they are athero-susceptible and the females are infertile. Notably, HDL from Scarb1^−/− mice is free cholesterol (FC)-rich. These observations support our hypotheses that the metabolic abnormalities associated with high plasma HDL levels are due to high HDL-FC bioavailability, and that reduction of HDL-FC bioavailability prevents/reverses these abnormalities. Our hypothesis is supported by other observations. FC transfer from HDL is fast, t = <5 min and reversible; in mice the majority of HDL- and nascent (n)HDL-FC transfers to the liver with t_1/2 <5 min without esterification; over longer times, nHDL-derived FC appears in all tissues (34). Whereas FC-free recombinant HDL (rHDL) supports cellular FC efflux, rHDL containing ≥15 mol% FC drives FC influx; the magnitude of rHDL-FC influx is proportional to HDL-FC content (35). Bacterial serum opacity factor (SOF) disrupts HDL structure, and lowers plasma total HDL levels in mice (36–38). Adeno-associated virus delivery of SOF to Scarb1^−/− mice reduces HDL-FC to nil and constitutively rescues fertility among the female Scarb1^−/− mice within ~24 h. In conclusion, SOF-mediated reduction of HDL particle number reduces HDL-FC bioavailability and could reduce metabolic problems due to high plasma HDL levels.

2. Pascal Bernatchez, Ph.D.

Associate Professor, Department of Anesthesiology, Pharmacology & Therapeutics, University of British Columbia, Heart & Lung Innovation Centre, St. Paul’s Hospital, Vancouver, British Columbia, Canada. HDL, LDL and Muscular Dystrophy.

Lipid management pharmacotherapy has highlighted new links between LDL cholesterol-lowering approaches and skeletal muscle homeostasis (39). For instance, statins can in rare cases cause rhabdomyolysis, but how this occurs is poorly understood. Using genetic models of muscular dystrophy (MD) that exhibit weak but chronic muscle degeneration (Duchenne and Limb-Girdle MD type 2b mice), we have observed drastically worsened skeletal muscle wasting and even loss of ambulation when non-HDL-C levels are increased via APOE gene inactivation and high fat diets (40). Double Duchenne/apoE or Limb-Girdle 2b/apoE mice can show severe muscle wasting compared to WT, apoE or single MD mice, with muscle lesion area jumping from 1–5% in skeletal muscles of Duchenne or Limb-Girdle 2b mice to up to 64–94% in double mutant mice (41). More interesting is the fact that muscle damage can reach levels that cause complete loss of ambulation, a phenotype never reported in these mice. Together these data clearly showed an unexpected linked between plasma lipoprotein levels and muscle homeostasis in already genetically weakened tissues and even suggested that non-HDL-C may possess intrinsic toxic properties in muscles. In various settings of Duchenne MD, plasma lipoprotein metabolism was analyzed, and HDL-related abnormalities were also noted even in settings of low muscle damage as assessed by normal creatine kinase levels. Together our data suggest a profound defect in LDL- and HDL-associated cholesterol metabolism in MD, which warrants further investigation.

3. Amanda Ribeiro Martins da Silva, Ph.D. graduate student

(Dr. Graziella E. Ronsein, mentor), Institute of Chemistry, University of Sao Paulo, Sao Paulo, Brazil. High-density lipoproteins subclasses mapping by targeted quantitative proteomics.

HDL is a heterogeneous particle and linked to a variety of cardio-protective functions (42). Close to 100 proteins have been identified as associated with HDL. However, apolipoprotein A-I (APOA1) and apolipoprotein A-II (APOA2), the two most abundant proteins in lipoprotein, make up to 90% of the protein content (43). Thus, the proteins may be differentially localized to distinct HDL particles, and most of them are present in low abundance, which makes accurate quantification a challenge (44). Mass spectrometry (MS)-based targeted quantitative proteomics is a powerful approach to accurately differentiate HDL subclasses due to its high sensitivity, specificity, and wide dynamic range (45). We therefore quantified the proteome of HDL subclasses of 19 apparently healthy volunteers. HDL particles were separated into HDL₂ and HDL₃ subfractions by discontinuous density ultracentrifugation and their proteome was quantified employing two targeted methodologies, Data Independent Acquisition (DIA) and Parallel Reaction Monitoring (PRM). Both methods worked equally well with high precision to differentiate HDL subclasses. The majority of proteins were shared by both subclasses, but their abundance varied considerably. Dense HDL₃ was significantly enriched with proteins related to antioxidant activity, such paraoxonase and apoliprotein L1, apolipoprotein D, apoliprotein A-IV, apolipoprotein H, apolipoprotein J and phosphatidylcholine-sterol acyltransferase. While HDL₂ was enriched with APOA2, apoliprotein Cs (apoC-I, apoC-II, apoC-III), apolipoprotein E, and serum amyloid A-4. Thus, DIA and PRM are attractive strategies to differentiate HDL subclasses proteome and may contribute to understand how HDL proteome affects its functionality.

Session Five: Microbiota and HDL Metabolism

1. Jacob L. Barber, Ph.D. graduate student

(Dr. Mark A. Sarzynski, mentor). Department of Exercise Science, University of South Carolina, Columbia, South Carolina. Association of exercise-induced changes in cholesterol efflux capacity with changes in the HDL proteome.

HDL function has recently been identified as a strong predictor of cardiovascular risk (32). Exercise is known to effect HDL metabolism and cardiovascular disease risk, however the effects of exercise on measures of HDL function and composition are less well known (46–48). We examined changes in cholesterol efflux capacity (CEC) and concomitant changes in the HDL proteome in 19 white, non-related individuals from the HERITAGE Family Study. Measurement of the efflux of radiolabeled and BODIPY-labeled cholesterol from J774 macrophages to apolipoprotein B depleted plasma, along with untargeted HDL proteome profiling, was performed at baseline and following 20-weeks of endurance exercise. There was no effect of regular exercise on CEC; however, exercise training resulted in improvements in CEC in 9 subjects (responders) and either no changes or decreased CEC in 10 subjects (non-responders). A total of 33 known HDL proteins were identified in all 19 subjects. There were no overall changes in abundance of any HDL proteins following training. Changes in abundance of six HDL proteins (Albumin, IGHA1, IGHG2, C6, SAA4, and APOE) were nominally negatively correlated with changes in CEC. Additionally, among responders HDL APOE abundance was significantly decreased (p=0.03), while there was no change in HDL APOE abundance (p=0.78) among non-responders. This pilot study provided limited evidence for a potential role of select HDL proteins as mediators of changes in CEC. Future studies with larger sample sizes are needed to further examine how alterations in the HDL proteome are related to HDL function.

2. Gregory C. Shearer, Ph.D.

Associate Professor, Nutritional Sciences, The Pennsylvania State University, University Park, Pennsylvania. Eicosapentaenoate and other polyunsaturated fatty acids in HDL function.

Recent clinical trials demonstrate 4 g/d of eicosapentaenoic acid (EPA) is associated with a 25% risk reduction for myocardial infarction (49). Further, subjects with low HDL had a greater benefit from EPA than those with high HDL. EPA is an ω3-polyunsaturated fatty acid (PUFA). There are strong reasons to associate its effects with HDL: HDL transport PUFAs; HDL responds most to ω3-PUFA treatment; and HDL host more oxylipins than any other plasma pool (oxylipins are signaling metabolites derived from PUFAs, including EPA) (50–51). Increased ω3-PUFA intake is associated with large increases in HDL-EPA-oxylipins. We seek to explain PUFA action in general, and EPA action specifically, by testing whether HDL mediates the cellular efflux of oxylipins. Oxylipins are intracellular oxylipins and by acting as an oxylipin acceptor, HDL could regulate intracellular signaling. For detection purposes, we used 12-HETE as an oxylipin tracer. It is produced by 12-lipoxygenase from arachidonate (AA). We hypothesized efflux of 12-HETE from activated macrophages would be apoA-I and ABCA1 dependent. RAW264.7 macrophages were provided d8-AA and stimulated with 100 ng/mL LPS under four conditions: ±apoA-I, and ±ABCA1 expression by siRNA silencing. Tracer efflux was analyzed using compartmental modeling. The efflux of label into media phospholipids (as media esterified) was dependent on the presence of both apoA-I and ABCA1 and concurrently resulted in the lowest intracellular unesterified d8–12-HETE. Oppositely, the absence of apoA-I or ABCA1 resulted in high intracellular label appearance and little to no appearance of d8–12-HETE in the media phospholipid pool. Overall, our results suggest an important role of HDL in mediating oxylipin efflux.

3. Angela M. Zivkovic, Ph.D.

Assistant Professor, Department of Nutrition, University California Davis, Davis, CA. Diet-induced changes in HDL lipidome, glycoproteome, and functional capacity.

There is growing evidence that the composition and functional capacity of HDL particles are important determinants of the protective effects of HDL in cardiovascular disease. But we do not yet know how to improve the composition and function of HDL. In two controlled, randomized-order, cross-over feeding studies we determined the effects of short-term diet and long-term diet on HDL composition and function. Four days of fast food vs. Mediterranean diet widely remodeled the HDL lipidome to reflect dietary lipid composition, altered apoA-I content but not A1AT, A2HSG, apoC-III, apoE, SAA, and did not alter protein glycoprofiles, except apoC-III sialylation. Individual HDL lipid species and glycopeptides were highly correlated with cholesterol efflux capacity; however, the diet-induced compositional changes were not accompanied by changes in HDL CEC possibly due to differences in baseline total:HDL cholesterol ratios, study duration, or gut microbiome composition. Four weeks of two whole eggs vs. egg whites per day, on the other hand, increased HDL CEC, which was unrelated to HDL ApoA-I content, particle size or distribution, or lipidomic compositional changes (52). Our findings indicate that diet can alter HDL composition in as little as 4 days. High inter-individual variability in response even in a tightly controlled study, may be partly explained by baseline lipid profiles and gut microbiome. Our results also indicate that diet can alter HDL CEC, and that HDL lipids and glycoproteins are highly correlated with HDL function, providing evidence that dietary strategies are a promising approach for improving the functional capacity of HDL.

4. Uwe J.F. Tietge, M.D., Ph.D.

Professor, Division of Clinical Chemistry, Karolinska Institutet, Stockholm, Sweden. Impact of intestinal microbiota on reverse cholesterol transport.

Mediating RCT is envisioned as a major anti-atherosclerotic function of HDL particles. Available data lend strong support to a concept, in which, at least in rodents, the biliary secretion pathway accounts for approximately 75% of RCT while the remaining 25% occurs from the non-biliary pathway of trans-intestinal cholesterol excretion (53–54). Within the biliary pathway, cholesterol can be secreted either directly or after metabolic conversion into bile acids. RCT studies distinguishing between the respective excretion of the macrophage-derived cholesterol tracer within the fecal neutral sterol versus the fecal bile acid fraction demonstrated that a large proportion of RCT actually occurs via bile acids (55). While primary bile acids are formed in the liver, secondary bile acids are metabolic products of the intestinal microbiota. Germ-free mice lack intestinal bacteria and thus the capacity for producing secondary bile acids. Interestingly, in the absence of a microbiota RCT is increased due to more secretion of RCT-relevant cholesterol within the fecal bile acid fraction (56). Since mostly the rodent-specific muri-cholic acids were affected, more research in the human system is required to investigate, if such results can be translated. Nevertheless, bile acid sequestrants represent a classical successful anti-atherosclerotic intervention strategy with relatively limited side effects. Conceivably, with the application of bile acid sequestrants RCT within the fecal bile acid fraction could be enhanced. As a secondary effect also RCT within the fecal neutral sterol fraction could increase, since the availability of bile acids is important also for cholesterol absorption.

Jack Oram Research Award

Godfrey S. Getz, M.D., Ph.D., Chairman Emeritus, Donald N. Pritzker Distinguished Professor, Departments Pathology, Biochemistry, Molecular Biology, University of Chicago, Chicago, Illinois. Apolipoprotein A-I and HDL heterogeneity in humans and mice.

Dr. Getz focused on three topics. The first being differences between HDL and apoA-I in humans and mice, based upon the role of apoA-I in distinguishing between HDL₂ (mouse only) and HDL₃ and HDL₂ in human plasma. Chimeric apoA-I containing mouse helices 7 and 8 in a human apoA-I background favored HDL₂ distribution. Second, it was noted that despite the strong evidence for atheroprotection by apoA-I derived from animal experiments, the APOA1 gene does not appear as a causal gene from genome wide association studies in humans and the Hybrid Diversity Mouse Panel (HMDP). This may be reconciled by the possibility that apoA-I serves as a platform protein for HDL formation and/or the compositional heterogeneity of HDL. An example of the latter is the finding that plasminogen promotes ABCA1 dependent cholesterol efflux (57). Plasminogen is found in a small cholesterol free phospholipid containing particles. The fact that apoA-I does not appear as a major gene in HMDP is exemplified by the comparison of C57BL6 and FVB mice, atherosensitive and atheroresistant strains, respectively. The latter strain shows limited or no effect of apoA-I deficiency on atherosclerosis. The comparison of HDLs from the two strains shows a polymorphism of both apoA-I (2 amino acid differences) and apoA-II (3 amino acid differences). HDL of FVB mice is enriched in apoA-II. The latter protein of the two strains show differences in in vivo dynamics. HDL can be remodeled by exposure to tandem mimetic peptides; FVB HDL is much more stable to this remodeling. Advantage could be taken of this remodeling, combined with proteomics could be used to probe the structure of HDL by examining the release of other HDL proteins along with apoA-I. Lastly, a novel method of in vivo cholesterol efflux was described, involving the encapsulation of macrophages in an alginate capsule, allowing for the recovery of the macrophages for assessment of sterol homeostasis and gene expression in the macrophage subject to various global environments. The method could also be used to encapsulate other cells or two different cell types for their direct interaction at the level of gene expression.

Perspectives

HDL exists. A simple statement and a challenge for basic scientists and clinicians to now convincingly determine for what biological purpose(s).

As evidenced by the talented roster of established investigators and emerging ones from a variety of scientific backgrounds, there is a vibrant community of scientists exploring new paradigms of HDL biology. We are continuing to discover novel functions of HDL particles, including for example, their role in protection from toxin-induced damage via SR-B1, their role in enhancing lymphatic transport, and their role as transporters/effluxers of lipid signaling molecules (i.e. oxylipins). An area of continued focus was on the composition of HDL-p, from different methodologies to examine their number, size, proteome, and association with functionality that largely is still defined by cholesterol efflux assays. Environmental effects on the HDL proteome were examined by interventions with exercise and diet, with intriguing early results that need testing in larger populations. From a clinical perspective, results from the JUPITER trial suggest that measurement of HDL particles vs. cholesterol might be more informative in CVD risk reduction. All of this speaks to an underlying friction in the need to distinguish the clinical utility of HDL-p vs. the lipid content of heterogenous HDL molecules.

Looking forward, the 10^th HDL Workshop will take place in May 2020 in Chicago. The chair will be Dr. Annabelle Rodriguez and the co-chair will be Dr. Kasey C. Vickers. The roster of speakers will continue to be innovative established investigators and emerging ones, with topics that will push the boundaries of what is known about HDL biology. The tentative program will include topics on HDL and reproduction, HDL as a drug platform, clinical perspectives of high HDL-C paradox, non-traditional HDL cargo, and HDL immunology.

We believe that health care organizations such as the American Heart Association and others should issue treatment guidelines that specifically address the role of HDL-C in the clinical management of patients at risk and those with CVD. The overwhelming evidence shows that low HDL-C adds predictive value for increased CVD risk while high HDL-C does not show evidence for cardioprotection. What about the question of HDL particles and functionality? Results from large clinical trials do point to the possibility of HDL particles and cholesterol efflux assays as providing utility in CVD risk reduction. However, much more work needs to be done to determine if these assays can be performed in a high throughput manner relevant to clinical medicine and insurance reimbursement.

We believe it is important to expand into other fundamental aspects of HDL biology, as shown by our program development for the 10^th HDL Workshop. These explorations will likely guide us into causal pathways influenced directly or secondarily by HDL that will be targets for future clinical diagnostic and therapeutic applications.

The clinical need is urgent and this collaborative effort of academic basic scientists, clinicians, and industry researches will certainly find the answers to why HDL exists.

Highlights.

• HDL Workshop celebrates its 9^th year as a forum for academic basic scientists, clinical investigators, and industry researchers to have candid discussions about HDL biology

• Presenters discussed novel aspects of scavenger receptors on cardiomyocyte function, LDL transendothelial cytosis, and lipid raft distribution.

• Presenters discussed many methodologies to evaluate the HDL proteome and association with cardiovascular disease.

• Presenters discussed influences like serum opacity factor, omega-3 fatty acids, egg consumption, exercise, and the intestinal microbiome on different aspects of HDL biology.

• Dr. Godfrey S. Getz provided a comprehensive discussion about apoA-I and HDL proteome in atherosclerosis as well as new in vivo approaches to assess HDL functionality.

Nonstandard abbreviations

CVD

Cardiovascular disease

SOF

serum opacity factor