The Role of High Density Lipoprotein Compositional and Functional Heterogeneity in Metabolic Disease

The role of high density lipoprotein compositional and functional heterogeneity in metabolic disease

Scott M. Gordon

B.S. State University of New York College at Brockport

October, 2012

A Dissertation Presented to the Faculty of The University of Cincinnati College of Medicine in partial fulfillment of the requirements for the Degree of Doctor of Philosophy from the Pathobiology and Molecular Medicine graduate program

W. Sean Davidson Ph.D. (Chair) David Askew Ph.D. Professor and Thesis Chair Professor Department of Pathology Department of Pathology University of Cincinnati University of Cincinnati

Francis McCormack M.D. Gangani Silva Ph.D. Professor Assistant Professor Department of Pathology Department of Pathology University of Cincinnati University of Cincinnati

Jason Lu Ph.D. Assistant Professor Division of Bioinformatics Cincinnati Children’s Hospital

Abstract

High density lipoproteins (HDL) are complexes of phospholipid, cholesterol and protein that circulate in the blood. Epidemiological studies have demonstrated a strong inverse correlation between plasma levels of HDL associated cholesterol (HDL-C) and the incidence of cardiovascular disease (CVD). Clinically, HDL-C is often measured and used in combination with low density lipoprotein cholesterol (LDL-C) to assess overall cardiovascular health. HDL have been shown to possess a wide variety of functional attributes which likely contribute to this protection including anti-inflammatory and antioxidative properties and the ability to remove excess cholesterol from peripheral tissues and deliver it to the liver for excretion, a process known as reverse cholesterol transport. This functional diversity might be explained by the complexity of HDL composition. Recent studies have taken advantage of advances in mass spectrometry technologies to characterize the proteome of total HDL finding that over 50 different proteins can associate with these particles. This adds to a growing body of evidence that supports the global hypothesis of this thesis which is that the total pool of HDL in an individual is composed of numerous subspecies with distinct protein and lipid compositions and therefore will have distinct functional properties. Additionally, we believe that the composition of HDL is dynamic and can change in response to changes in the environment of the blood, as can occur in disease. To test these hypotheses we devised an approach based on three aims. Aim 1: Identify and characterize HDL subspecies based on protein composition. Aim 2: Analyze functional heterogeneity across separated plasma HDL fractions. Aim 3: Examine the effect of type 2 diabetes on HDL subspecies distribution in young adults. To accomplish these goals we have developed novel methods for the separation and fractionation of HDL subspecies from human plasma and their subsequent proteomic and functional analysis. Our findings support the existence of a diverse array of HDL species with distinct functionalities which correlate strongly with specific lipid or protein components of the particle. Additionally, we found that the composition of specific subfractions of HDL are altered in type 2 diabetes and that these fractions can predict vascular health better than LDL-C or HDL-C, suggesting that these may prove to be better biomarkers for cardiovascular risk than the current clinical standard. A detailed understanding of HDL subspeciation will be invaluable in the further development of HDL as a biomarker and therapeutic target not only for cardiovascular disease but a variety of disease states.

This dissertation is based, in part, on the following published manuscripts. Permission to include the information contained within these manuscripts was obtained from the publishers (Appendix 1).

1. Gordon, S., Durairaj, A., Lu, J., and Davidson, W.S. HDL Proteomics: Identifying New Drug Targets and Biomarkers by Understanding Functionality. (2010) Cur. Cardiovasc. Risk Reports., 4: 1-8.

2. Gordon, S.M., Jingyuan Deng, L. Jason Lu, and W. Sean Davidson. Proteomic characterization of human plasma high density lipoprotein fractionated by gel filtration chromatography. (2010) J. Proteome Res., Oct. 1; 9(10):5239-49.

3. Gordon, S.M., Hofmann, S., Askew, D., Davidson, W.S. High Density Lipoprotein: It’s Not Just About Lipid Transport Anymore. (2011) Trends in Endocrinology and Metabolism. 22, 9-15.

iii

Acknowledgements

This work was supported by an American Heart Association Pre-doctoral Fellowship, a University Research Council Graduate Student Research Fellowship from the University of Cincinnati and by a pilot grant from the Cincinnati Diabetes and Obesity Center.

I want to thank my thesis advisor Sean Davidson for his outstanding guidance and encouragement and for helping me to obtain the best possible opportunities to develop as a successful independent scientist.

Adam Rich: My undergraduate research advisor who introduced me to the world of laboratory research and the scientific community. I cannot express enough appreciation for the way you took me in and allowed me to think and grow in your lab.

Kathy and Phil Gordon: My amazing parents who have always encouraged me to reach for the stars and told me that there is nothing I cannot accomplish if I put my mind to it and work hard.

Table of Contents

Abstract……………………………………………………………………..………… ii

List of Figures………………………………………………………………………… vii

List of Tables…………………………………………………………………………. ix

Chapter 1. Introduction and background

A. High density lipoproteins 1. Clinical importance of HDL………………………………….. 1 2. Functions of HDL…….…………………………………...….. 4 3. Current measures of HDL subfractions……………….…... 17 4. Applications of modern proteomics to HDL…….……..…... 20 5. Evidence for HDL subspeciation based on protein content ……………………………………………………..…. 24 6. HDL subspeciation and its impact on the treatment of cardiovascular disease……………………………………... 27 B. Goals of thesis research 1. Aim 1: To identify and characterize HDL subspecies based on protein composition…………………………….. 30 2. Aim 2: Analyze functional heterogeneity across separated plasma HDL fractions…………………………. 31 3. Aim 3: Examine the effect of type 2 diabetes on HDL subspecies distribution in young adults…………………… 31

Chapter 2. Proteomic characterization of human plasma high density lipoprotein fractionated by gel filtration chromatography

A. Introduction……………………………………………….…..………. 57 B. Experimental………………………………………………….………. 60 C. Results……………………………………………………..…….……. 63 D. Discussion…………………………………………………………….. 69 E. Conclusions…………………………………………………………… 74

Chapter 3. A novel correlation analysis to determine HDL subspecies protein composition

A. Introduction……………………………………………………..…….. 91 B. Experimental………………………………………………………….. 92 C. Results………………………………………………………………… 95 D. Discussion……………………………………………………………. 99

Chapter 4. Using apolipoprotein deficient systems to study HDL composition

A. Introduction…………………………………………………………… 114 B. Experimental………………………………………………………….. 116 C. Results………………………………………………………………… 117 D. Discussion…………………………………………………………….. 121

Chapter 5. Functional analysis of HDL subfractions

A. Introduction…………………………………………………………... 145 B. Methods…………………..…………………………………………... 146 C. Results………………………………………………………………… 148 D. Discussion……………………………………………………………. 152

Chapter 6. Effects of type 2 diabetes on lipoprotein composition and arterial stiffness in adolescents and young adults

A. Introduction…………………………………………………………... 162 B. Methods…………………..…………………………………………… 164 C. Results………………………………………………………………… 168 D. Discussion……………………………………………………………. 174

Chapter 7. Discussion

A. Thesis Summary..………………………………………………….... 200 B. Contributions of this work.………………………………………….. 203 C. The future of HDL subspecies research and clinical implications. 207

Appendix I. Reproduction Rights…………………………………………………. x Appendix II. Supplemental Data………………………………………………..…. xiii

List of Figures

Chapter 1.

1-1. The increasing functional heterogeneity of high density lipoprotein.

Chapter 2.

2-1. Elution profiles from Superose 6 (2x) and Superdex 200 (3x) size exclusion chromatography configurations. 2-2. SDS PAGE comparison of total HDL preparations derived from ultracentrifugation (UC) and gel filtration (GF) chromatography. 2-3. Ability of calcium silicate hydrate (CSH) to bind ultracentrifugally isolated human plasma lipoproteins. 2-4. Ability of CSH to bind phospholipid-containing particles from fractions collected by gel filtration chromatography. 2-5. Examples of elution profile shifts for proteins upon ether delipidation of fresh human plasma. 2-6. Lipid-associated proteins identified in the plasmas of 3 normolipidemic donors. 2-7. Distribution patterns of common HDL associated proteins across gel filtration fractions. 2-8. Triple Superdex distribution profiles for identified lipid- associated proteins. 2-9. Gene Ontology functional associations of newly identified lipoprotein associated proteins.

Chapter 3.

3-1. Correlation strategy for the identification of phospholipid particle subspecies. 3-2. Lipid and protein distribution profiles produced by gel filtration chromatography. 3-3. Lipid and protein distribution profiles produced by anion exchange chromatography. 3-4. Lipid and protein distribution profiles produced by isoelectric focusing. 3-5 Heat map displaying protein distribution patterns by anion exchange chromatography. 3-6. Heat map displaying protein distribution patterns by isoelectric focusing. 3-7. Tracking co-migratory patterns of protein pairs across different separation techniques.2-14. Lipid distribution profiles for apolipoprotein knockout mice.

vii

Chapter 4.

4-1. Lipid distribution profiles for apoA-I knockout mice. 4-2. Lipid distribution profiles for apoA-II knockout mice. 4-3. Lipid distribution profiles for apoA-IV knockout mice. 4-4. Lipid distribution profiles for apolipoprotein knockout mice. 4-5. Strategy for identifying proteins affected in knockout models. 4-6. Characterization of apoA-I knockout mice. 4-7. Distribution profiles of some proteins shift as a result of apoA-I knockout. 4-8. Distribution profiles of some proteins shift as a result of apoA-II knockout. 4-9. Distribution profiles of some proteins shift as a result of apoA-IV knockout. 4-10. Number of protein shifts detected in each knockout model. 4-11. Phospholipid distribution profiles for control and apoA-I deficient human plasma fractions collected by gel filtration chromatography. 4-12. Venn diagram displaying total identified protein numbers between control and A-I deficient human plasma fractions. 4-13. Comparison of protein distribution profiles.

Chapter 5.

5-1. Cholesterol efflux capacity across gel filtration fractions. 5-2. Anti-oxidative capacity across gel filtration fractions. 5-3. Anti-inflammatory capacity across gel filtration fractions. 5-4. Summary of functional distribution profiles.

Chapter 6.

6-1. Lipid distribution profiles are altered in type 2 diabetes. 6-2. Correlations of standard clinical lipid measures with pulse wave velocity. 6-3. Lipoprotein fractions correlate with pulse wave velocity. 6-4. Scatterplots for the fractions with the strongest observed correlation with pulse wave velocity. 6-5. Adjustment of proteomics data. 6-6. Distribution patterns of phospholipid associated proteins are altered in type 2 diabetes.

Chapter 7.

7-1. Model for association of protein components with HDL. 7-2. Defining lipoproteins based on primary lipid components.

viii

List of Tables

Chapter 1.

Table 1-1. Master compendium of HDL associated proteins identified using modern proteomic strategies.

Chapter 2.

Table 2-1. Quantitative binding of ultracentrifugally isolated LDL and HDL lipids by LRA.

Chapter 3.

Table 3-1. Results of correlation analysis.

Chapter 4.

Table 4-1. Proteins with shifted distribution profiles in different knockout models.

Table 4-2. List of all identified proteins and whether or not they were shifted in apoA-I deficiency.

Table 4-3. Correlation analysis results supported by mouse knockout studies.

Chapter 5.

Table 5-1. Heat map of protein distribution and functional correlation analysis.

Chapter 6.

Table 6-1. Clinical characteristics of study participants. Table 6-2. Comparison of phospholipid associated proteins between groups. Table 6-3. Differences in protein distribution patterns between groups.

Appendix I. Reproduction Rights

Gordon, S., Durairaj, A., Lu, J., and Davidson, W.S. HDL Proteomics: Identifying New Drug Targets and Biomarkers by Understanding Functionality. (2010) Cur. Cardiovasc. Risk Reports., 4: 1-8.

SPRINGER LICENSE TERMS AND CONDITIONS Aug 13, 2012 ______

This is a License Agreement between Scott M Gordon ("You") and Springer ("Springer") provided by Copyright Clearance Center ("CCC"). The license consists of your order details, the terms and conditions provided by Springer, and the payment terms and conditions. All payments must be made in full to CCC. For payment instructions, please see information listed at the bottom of this form. License Number 2967190806070 License date Aug 13, 2012 Licensed content publisher Springer Licensed content publication Current Cardiovascular Risk Reports Licensed content title High-Density Lipoprotein Proteomics: Identifying New Drug Targets and Biomarkers by Understanding Functionality Licensed content author Scott Gordon Licensed content date Jan 1, 2009 Volume number 4 Issue number 1 Type of Use Thesis/Dissertation Portion Full text Number of copies 5 Author of this Springer article Yes and you are the sole author of the new work Order reference number Title of your thesis / dissertation Proteomic and Functional Characterization of High Density Lipoprotein Subspecies Expected completion date Nov 2012 Estimated size(pages) 175 Total 0.00 USD

Gordon, S.M., Jingyuan Deng, L. Jason Lu, and W. Sean Davidson. Proteomic characterization of human plasma high density lipoprotein fractionated by gel filtration chromatography. (2010) J. Proteome Res., Oct. 1; 9(10):5239-49.

Gordon, S.M., Hofmann, S., Askew, D., Davidson, W.S. High Density Lipoprotein: It’s Not Just About Lipid Transport Anymore. (2011) Trends in Endocrinology and Metabolism. 22, 9-15.

ELSEVIER LICENSE TERMS AND CONDITIONS Aug 13, 2012 ______

This is a License Agreement between Scott M Gordon ("You") and Elsevier ("Elsevier") provided by Copyright Clearance Center ("CCC"). The license consists of your order details, the terms and conditions provided by Elsevier, and the payment terms and conditions. All payments must be made in full to CCC. For payment instructions, please see information listed at the bottom of this form. Supplier Elsevier Limited The Boulevard,Langford Lane Kidlington,Oxford,OX5 1GB,UK Registered Company Number 1982084 Customer name Scott M Gordon Customer address University of Cincinnati

Cincinnati, OH 45237 License number 2967190434920 License date Aug 13, 2012 Licensed content publisher Elsevier Licensed content publication Trends in Endocrinology & Metabolism Licensed content title High density lipoprotein: it's not just about lipid transport anymore Licensed content author Scott M. Gordon,Susanna Hofmann,David S. Askew,W. Sean Davidson Licensed content date January 2011 Licensed content volume number 22

xii

Appendix 2: Chapter 2 Supplemental Data

Supplement Table 2-1. Proteins identified in ultracentrifugally isolated HDL prepared for MS analysis by either chloroform-methanol lipid extraction or by calcium silicate hydrate procedure. HDL isolated from plasma by ultracentrifugation was subjected to either a chloroform-methanol delipidation (Lipid ext.) or the calcium silicate hydrate procedure (CSH), described in the Experimental section, prior to MS analysis. Data are displayed as number of spectral counts. While the spectral counts for the highly abundant apoA-I were lower for the CSH technique, the lower abundance proteins were detected equally well, or in some cases better, by the instrument.

xiii

Supplement Table 2-2. Proteins found to bind to CSH by mechanisms that are independent of phospholipid. The listed proteins were found in fractions that contain lipoproteins but failed to undergo a detectable elution pattern shift in response to prior plasma ether lipid extraction. This suggests that these proteins are not normally bound to lipids in plasma.

Identified proteins: Alpha-1B-glycoprotein Ceruloplasmin Coagulation factor X Coagulation factor XII Complement component C6 Complement component C7 Complement component C8 alpha chain Complement component C8 beta chain Complement component C9 Complement factor H Complement factor I Fibrinogen beta chain Fibrinogen gamma chain Haptoglobin Histidine-rich glycoprotein Ig alpha-1 chain C region Ig gamma-1 chain C region Ig lambda chain V-III region LOI Ig mu chain C region N-acetylmuramoyl-L-alanine amidase Plasma kallikrein Plasminogen Serum albumin Vitamin D-binding protein

xiv

Supplement Figure 2-1. Protein-Protein Interactions between newly identified lipid associated proteins and known HDL associated proteins. In this PPI network, black circles denote known HDL proteins, black triangle nodes denote 14 new phospholipid associated proteins. The interactions between known HDL proteins and new phospholipid associated proteins are in bold dash lines. The detailed interaction information between known HDL proteins and new HDL proteins are listed in the following table (one-step interaction also included).

New HDL proteins Between Known HDL proteins

C2 C3

C5 C3

CFB C3

C2 C4B

CFB C4B

SERPING1 F2

SERPINC1 F2

SERPIND1 F2

C1QB MYOC A2M

C1QB HRG FGA

C2 C3 GC

C5 CPN1 C3

C5 C2 C4B

C5 C7 CLU

C5 C8B CLU

C5 C6 F2

C5 C3 GC

C5 CPN1 KNG1

CFB C3 GC

CLEC3B PLG KNG1

CLEC3B PLG SERPINF2

IGFALS IGFBP5 VTN

SERPINA3 CELA1 A2M

SERPINA3 KLK2 A2M

xvi

SERPINA3 KLK3 A2M

SERPINA3 CTSG C3

SERPINA3 APP CLU

SERPINA3 F5 F2

SERPINA3 DNAJC1 ITIH4

SERPINA3 APP KNG1

SERPINA3 ADAMTS4 SERPINA1

SERPINA3 KLK3 SERPINA1

SERPINA3 KLK2 SERPINA4

SERPINA3 ELANE SERPINF2

SERPINA3 KLK2 SERPINF2

SERPINC1 KLK2 A2M

SERPINC1 F2 FGA

SERPINC1 KLK2 SERPINA4

SERPINC1 KLK2 SERPINF2

SERPINC1 KLK6 SERPINF2

SERPIND1 CTSG C3

SERPIND1 F2 FGA

SERPIND1 ELANE SERPINF2

SERPINF1 MYOC A2M

SERPINF1 CSNK2A1 VTN

SERPING1 ELANE KNG1

SERPING1 F2 KNG1

SERPING1 ELANE SERPINA1

xvii

SERPING1 PLG SERPINF2

xviii

Chapter 1: Introduction and background

Clinical importance of HDL

High density lipoproteins (HDL) are a circulating, non-covalent assembly of amphipathic proteins (~50% by mass) that stabilize lipid emulsions composed of a phospholipid monolayer (PL, ~30%) embedded with free cholesterol (~5%), with a core of triglycerides (TG, ~3%) and cholesteryl esters (CE,~12%). Epidemiological studies dating back to the 1950’s have shown that low plasma high density lipoprotein associated cholesterol (HDL-C) levels are a powerful risk factor for the development of cardiovascular disease (CVD). The study of HDL as a marker for cardiovascular risk started with the observation that patients presenting with a recent coronary event or complication of atherosclerosis had reduced protein and cholesterol associated with their HDL 1. Later, larger prospective studies such as the Framingham Heart Study established the power of HDL-C to predict cardiovascular risk 2. Meta analyses of multiple long term studies across varied populations demonstrated the consistency of this observation and revealed that a decrease in plasma HDL-C levels of 1 mg/dL could translate to a 2% to 3% increased risk for coronary artery disease 3. A recent systematic review of studies assessing the relationship between HDL-C and cardiovascular risk found 58 prospective studies, of these 31 found significant inverse correlations between HDL-C and all cardiovascular disease (CVD) outcomes studied and 17 found inverse correlations for some of the CVD outcomes studied 4.

Cumulatively this data leaves little doubt that HDL is playing a protective role against the development of cardiovascular diseases.

The inverse relationship between HDL-C and risk for CVD holds up across multiple populations and even under conditions involving different disease states. This relationship is apparent in both men and women, but is stronger in women 2, however the association tends to diminish around the time of menopause 5. Currently accepted clinical standards suggest that plasma HDL-C levels below 40 mg/dL (men) and 50 mg/dL (women) put one at increased risk for CAD while levels above 60 mg/dL (men and women) are protective. Studies have also demonstrated the persistence of this relationship in elderly populations 6 and across various racial groups 7, 8. Low HDL-C remains an important predictor of cardiovascular risk in patients with hypertension 9, type 2 diabetes 10, artery stenting 11, and human immunodeficiency virus (HIV) infection

12. Additionally, low HDL-C plays a more strongly weighted role among other factors in the ability of metabolic syndrome to predict cardiovascular risk 13. Low HDL-C has also been linked to the development of other diseases including cancer 14 and neurological disorders 15.

Due to the undeniable association between plasma HDL-C levels and CVD and the subsequent advances in our understanding of HDL’s roles in lipid metabolism in the human body, HDL has become an important target for development as both a biomarker and therapeutic target for cardiovascular disease. Strategies exploiting HDL for protection against CVD have largely been aimed at raising plasma levels of HDL-C.

This approach is most intuitive given the vast body evidence correlating circulating HDL-

C with cardiovascular risk. However, those working toward these goals have

2 encountered several recent road blocks confounding the seemingly clear linkage between HDL-C and CVD, raising question about the clinical utility of HDL raising therapies. The first is the failure of several HDL raising strategies to reduce cardiovascular risk. Several major drug companies have spent billions of dollars developing their own versions of a class of compounds called cholesterol ester transfer protein (CETP) inhibitors. These compounds act by different mechanisms to block the transfer of cholesterol from HDL to triglyceride rich lipoproteins such as very low density lipoproteins (VLDL). This transfer is a critical event in reverse cholesterol transport

(RCT), a process by which excess cholesterol is removed from peripheral tissues and delivered to the liver for recycling or conversion to bile acids for excretion via the feces.

Pharmaceutical companies have successfully developed such compounds, capable of raising HDL-C by 30 to over 100%. However, although these compounds effectively raise circulating HDL-C they have not successfully demonstrated protection against

CVD. Roche’s CETP inhibitor Dalcetrapib raised HDL-C by about 30% but was recently pulled from phase III clinical trials early due to lack of significant effect 16. Another HDL-

C raising compound nicotinic acid (niacin), effective raises HDL-C by about 21% and can reduce risk for CVD, however, this compound affects other clinical lipid measures as well, decreasing triglycerides by 15% and non-HDL-C by 7% making it unclear how much of the resulting benefit is due to elevation of HDL-C 17. The second road block raising question about the clinical utility of HDL was a large mendelian randomization study involving nearly 100,000 patients, which tracked naturally occurring genetic mutations that result in increased plasma HDL-C to determine if those with these mutations had lower incidence of myocardial infarction 18. The findings of this study

3 indicated that some genetic mechanisms that raise plasma HDL-C do not result in decreased risk for myocardial infarction.

We find ourselves in a confounding situation where epidemiological studies consistently demonstrate a significant inverse correlation between plasma levels of

HDL-C and the incidence of cardiovascular diseases, however, our efforts to utilize this relationship therapeutically have been largely unsuccessful. Even though we can successfully raise HDL-C levels, we are not rewarded with protection against CVD.

This must be telling us something about our understanding of HDL physiology… that it is more complicated than we thought. Our current clinical measure for HDL, HDL-C, may not be giving us all of the information we need to determine the impact of HDL on cardiovascular health. Further examination of the functionality and composition of HDL may provide a deeper understanding how HDL protects against CVD and why our current strategies for raising HDL are ineffective at preventing disease.

Functions of HDL

A widely accepted basis for the inverse relationship between human plasma HDL-C and

CVD is the ability of HDL, and its major protein constituent apolipoprotein (apo)A-I, to mediate RCT. The primary targets of this process are lipid-laden macrophages in the vessel wall, harbingers of the fatty streaks and atherosclerosis that can ultimately progress to myocardial infarction or stroke (for a recent review, see 19). A large body of evidence supports the notion that HDL-mediated RCT is essential for human cardiovascular health. In addition to numerous in vitro studies that demonstrate cellular

4 cholesterol efflux to HDL, there are well established biochemical pathways that process

HDL lipid via numerous circulating enzymes and transfer proteins as well as receptor- mediated uptake into the liver. The in vivo RCT assay developed by Rader and colleagues clearly shows that apoA-I over expression in mice promotes release of macrophage cholesterol first into the plasma compartment and then eventually to the feces, concomitant with a clear decrease in atherosclerosis susceptibility in this model

20. However, recent studies have provided tantalizing clues that HDL may be more than it appears. The application of mass spectrometry based proteomic approaches has revealed unexpected diversity in the HDL proteome (see section below). Interestingly, only about 1/3 of HDL proteins are known to mediate lipid transport. The rest play roles in such areas as protease inhibition, complement regulation and acute phase response.

This suggests that HDL has broader functions (Figure 1-1). On an evolutionary scale, atherosclerosis is a relatively recent affliction that strikes well after reproductive age and would not be expected to be a major driving force for apoA-I or HDL genetic evolution.

It follows that HDL probably evolved under selection pressure to support more basic survival functions.

Anti-inflammatory Functions. Aside from RCT, the next best recognized HDL function is its role as an anti-inflammatory regulator. It accomplishes this through interactions with both the vascular endothelium and circulating inflammatory cells. For example, HDL limits the extent to which endothelial cells can become activated by pro- inflammatory cytokines, resulting in reduced expression of adhesion molecules 21. This was elegantly demonstrated in vivo where infusion of rabbits with reconstituted HDL

(rHDL) reduced vascular inflammation by inhibiting endothelial cell adhesion molecule expression 22. The effect was diminished when the rHDL particle was generated with non-enzymatically glycated apoA-I 23 such as may occur in type II diabetes (T2D). HDL can also inhibit production of chemo-attractant molecules such as monocyte chemoattractant protein-1 (MCP-1) 24, a chemokine responsible for the recruitment of monocytes, dendritic cells and T lymphocytes to sites of injury and inflammation.

Additionally, HDL can modulate vascular tone by affecting the production of nitric oxide

(NO), a key mediator of vascular smooth muscle cell contraction. It accomplishes this by stimulating the activity of endothelial nitric oxide synthase (eNOS) to boost NO production, triggering vasorelaxation 25. Earlier in vitro work had demonstrated that apoA-I can increase eNOS activity via AMP-activated protein kinase (AMPK) activation

26, as well as the PI3K/AKT and MAP kinase signaling pathways 27. In vivo studies confirmed increased AKT and ERK1/2 phosphorylation in apoA-I transgenic animals and decreased AKT and ERK1/2 phosphorylation in aortas of apoA-I deficient mice 28.

Two cell surface proteins have also been implicated in HDL-mediated modulation of vascular tone, the scavenger receptor class B member 1 (SR-BI) and the ATP binding cassette transporter ABCG1. SR-BI co-localizes with eNOS in caveolae of vascular endothelial cells and interaction with HDL directly activates eNOS activity 29. ABCG1 is highly expressed in endothelial cells and is known to promote the efflux of cholesterol and oxysterols to HDL as well as to mediate various important intracellular processes

(reviewed in 30). It has been demonstrated that formation of eNOS dimers, which is necessary for proper enzyme function, is inhibited in aortas of ABCG1 deficient

(ABCG1-/-) mice fed high cholesterol or western diets 31. The oxysterol 7-

6 ketocholesterol tended to accumulate in aortic endothelial cells and femoral arteries from these mice displayed impaired vasorelaxation in response to the agonist acetylcholine. Treatment of WT aortic endothelial cells with HDL prevented 7- ketocholesterol induced disruption of eNOS dimer formation and restored eNOS activity, but this did not occur in ABCG1-/- cells. Furthermore, ABCG1-/- mouse endothelial cells displayed increased surface expression of inflammatory markers like E-selectin and

ICAM-1 and increased secretion of IL-6 and MCP-1 32. These findings correlated with increased endothelial cell activation in the ABCG1-/- mouse, suggesting that HDL- mediated lipid removal is an important mechanism to reduce vascular inflammation.

HDL also interacts directly with circulating leukocytes to limit inflammation. For example, apoA-I on HDL can inhibit activation of the monocyte cell surface protein

CD11b, an integrin involved in vascular adhesion 33. Interestingly, this process was dependent on another ATP binding cassette transporter, ABCA1. In this gene lies the molecular defects responsible for Tangier's disease, in which patients exhibit markedly depressed HDL levels. The transporter plays a critical role in HDL formation and cholesterol efflux via an interaction with lipid-free forms of apoA-I, which are either secreted after synthesis in the liver and intestine or displaced from intact lipoproteins.

The involvement of two ABC transporters with established cholesterol efflux capacity is strongly suggestive of a mechanistic link between the cholesterol efflux and anti- inflammatory properties of HDL. This connection has recently been elegantly demonstrated at the molecular level. The ABCA1 transporter is an important regulator of cholesterol efflux from lipid laden cells in the periphery to lipid-free apoA-I, a critical step in reverse cholesterol transport and HDL production. Interestingly, Oram and

7 colleagues have published a series of papers demonstrating that the apoA-I/ABCA1 interaction not only results in a hand off of cellular lipid to form HDL, but also mediates outside-in signaling events. One consequence is a triggering of enhanced apoA-I binding to ABCA1 mediated by Janus kinase 2 (JAK2) to reinforce lipid transfer 34.

More recently, a second arm of this pathway has been identified in which JAK2 activation by apoA-I binding to ABCA1 promotes the association of the STAT3 transcription factor with ABCA1 35. Once bound, STAT3 is phosphorylated to its activated state and translocates to the cell nucleus where it activates anti-inflammatory gene expression programs. An important result of this is a reduced inflammatory response to bacterial lipopolysaccaride. This discovery has solidified the relationship between the lipid transport functions of apoA-I and inflammatory response pathways in macrophages. It also raises the intriguing possibility that HDL or its components may be beneficial in the setting of other chronic inflammatory conditions such as rheumatoid arthritis. Finally, the binding of HDL to macrophage-derived cytokines and growth factors may limit the pro-inflammatory activity of these proteins. For example, progranulin is a pro-inflammatory protein secreted by macrophages that has been shown to bind to plasma apoA-I; incubation with apoA-I or HDL suppresses progranulin induced expression of inflammatory markers in HEK 293 cells 36.

Innate Immune Functions. The innate immune system represents the first line of defense against invading microorganisms. Accumulating evidence supports the idea that HDL are integral components of innate immunity, mediating diverse functions that defend against viral, bacterial and parasitic infections. The finding that HDL is host to a

8 number of complement factors reinforces this view and suggests that HDL acts as a platform for the assembly of potent immunomodulatory complexes that regulate antimicrobial activity 37, 38. In addition, since infection and inflammation are tightly linked processes, the ability of HDL to regulate the amplitude of the inflammatory response may work in conjunction with these direct antimicrobial effects to influence the outcome of the infection.

Bacterial - HDL are conserved in lower vertebrates such as fish, where they are expressed at high levels in plasma, as well as in tissues that constitute the primary defense barriers to bacterial infection 39. Recent studies have demonstrated that apoA-I has potent bactericidal and bacteriostatic effects against both gram-positive and gram- negative bacteria, and vaccination with bacterial preparations increased apoA-I levels and enhanced the antibacterial activity of serum 40. This antimicrobial effect is also effective against pathogens that infect humans, suggesting that HDL may have originally evolved as components of primitive innate immune systems 39, 41. In addition to their direct antibacterial effects, mammalian HDL have also been shown to protect against the adverse consequences of a bacterial infection by limiting the toxicity of bacterial components, many of which are responsible for life-threatening pathophysiological changes in the host. This toxin-neutralizing activity is largely attributed to apoA-I and has been shown to be effective against enterohemolysin 42, lipopolysaccharide (LPS) and lipoteichoic acid 43-47.

Viral - The antiviral effects of human serum have been known for some time and

HDL are known to account for a significant proportion of this activity, neutralizing both

DNA and RNA viruses as well as enveloped and non-enveloped viruses 48. The precise

9 mechanism for HDL-mediated antiviral activity is incompletely understood, but current evidence suggests that it can involve direct viral inactivation, interference with viral entry into the cell or inhibition of virus-induced cell fusion 48, 49.

Parasitic - The most well understood mechanism for the antimicrobial effects of

HDL involves Trypanosoma brucei, a eukaryotic parasite that is the causative agent of

African sleeping sickness 50. Humans are naturally resistant to infection by the subspecies T. brucei brucei because the parasite is highly susceptible to lysis by trypanosome lytic factor (TLF) in human serum 51. TLF is comprised of at least two independent circulating complexes: TLF1 and TLF2 52. TLF1 is bound to a subfraction of HDL and contains the primate-specific proteins apolipoprotein (apo)L-I and haptoglobin-related protein (Hpr) 51, 53. TLF2 contains the same two proteins, but is part of a lipid-poor complex comprised of IgM and apoA-I 54, 55. Within the circulation, the

Hpr component of TLF1 associates with hemoglobin (Hb), generating a Hb-charged

TLF1 particle that is efficiently taken up by the parasite due to a receptor-mediated endocytic process that trypanosomes use to acquire heme from the host bloodstream

56. Once inside the parasite, TLF1 is delivered to the lysosome by the endocytic pathway, where progressive acidification dissociates apoL-I from the complex, allowing it to insert into the lysosomal membrane. The bcl2-related pore-forming domain in the apoL-I protein triggers an influx of chloride ions into the lysosome, followed by water, resulting in swelling and trypanosome lysis. The pathway by which TLF2 enters the parasite is less clear, but the lysis mechanism appears to be the same. Interestingly, parasites that reside within the phagolysomes of macrophages, such as Leishmania sp,

10 are not protected from TLF because the infected macrophages deliver plasma TLF directly into the vacuole in which the Leishmania organisms reside 57.

If there is any doubt that HDL plays an important role in host defense, it may be of interest to consider some of the adaptations that invading pathogens have evolved to specifically target HDL or its components. For example, Streptococcus pyogenes, a group A streptococcal bacterium responsible for tonsilitis, pharyngitis, and toxic shock syndrome secretes a protein called serum opacity factor (SOF). Originally described for its ability to opacify the normally translucent plasma of infected individuals, SOF has recently been found to specifically target HDL particles by direct binding to apoA-I and apoA-II 58. This causes a dramatic redistribution of the HDL neutral lipid cargo into large protein poor micro-emulsions. Although the exact effect of this reaction on the competitive fitness of the bacterium has not yet been identified, it is likely that SOF evolved as a virulence factor designed to subvert the anti-bacterial properties of intact

HDL particles. Interestingly, parasites have also developed innovative strategies to frustrate HDL-based defenses. In contrast to T. b. brucei, humans can acquire African sleeping sickness by infection with the related subspecies T. b. rhodesiense. The ability of T. b. rhodesiense to infect humans has been attributed to the secretion of serum resistance-associated protein (SRA), a virulence factor that directly interacts with the C- terminal helix of HDL-associated apoL-I and inhibits its anti-trypanosomal activity 53.

However, variants of apoL-I have recently been identified in the African population that do not bind SRA and thus retain lytic activity against T. b. rhodesiense 59. These apoL-I variants are associated with high rates of renal disease in African Americans,

11 suggesting that the selection pressure to acquire an SRA-resistant variant of apoL-I in

Africa may have contributed to the development of a risk factor for kidney disease.

Modulation of Glucose Metabolism. Type II diabetes mellitus (T2D) is characterized by a lack of glucose control due to the development of insulin resistance. Patients with

T2D also display dyslipidemia with low HDL-C concentrations (reviewed in 60). It has been shown in cell culture, animal and human studies that apoA-I gene expression is decreased by elevated glucose levels and increased by insulin 61, but there is emerging evidence that HDL, and apoA-I in particular, may also directly modulate glucose metabolism. In ex vivo experiments, it was found that recombinant apoA-I improved glucose uptake associated with increased AMP-activated protein kinase (AMPK) activity in mouse skeletal muscle. It was also demonstrated that the absence of apoA-I in mice resulted in higher fasting blood-glucose levels associated with reduced AMPK activity and increased hepatic gluconeogenesis determined by increased PEPCK and G6Pase expression levels 62. Furthermore, apoA-I was internalized through a clathrin- dependent endocytotic process before activating AMPK and acetyl-coenzyme A carboxylase (ACC) in cell culture experiments. Recent observations indicate that infusions of recombinant HDL particles (rHDL) reduced plasma glucose, increased insulin secretion and promoted glucose uptake in skeletal muscle of patients with T2D

63. Based on cell culture experiments, the authors suggested that the AMPK signaling pathway may be involved in the rHDL-mediated increase in glucose uptake. Regarding the observed increased insulin levels in the T2D patients, the authors provided the first evidence that HDL directly stimulates insulin secretion from pancreatic beta cells using

12 a murine beta cell line. In a follow-up study using transformed beta cell lines and primary islets under basal and high-glucose conditions, it was determined that insulin secretion can be mediated by apoA-I and apoA-II, in the lipid-free form, as a constituent of rHDL or by HDL isolated from human plasma 64. Furthermore these effects were calcium dependent and involved expression of ABCA1, ABCG1, and SR-BI. The importance of ABCA1 in modulating insulin secretion has been underscored by the observation that mice with specific inactivation of ABCA1 in beta cells had markedly impaired glucose tolerance and defective insulin secretion but normal insulin sensitivity

65.

Besides its direct effect on glucose homeostasis, there is recent evidence that

HDL may improve insulin resistance and obesity though its anti-inflammatory actions.

The apoA-I mimetic peptide L-4F, known to have anti-oxidant properties, was investigated to determine if it would ameliorate insulin resistance and diabetes in genetically obese mice. It was found that L-4F treatment reduced adipocity, inflammatory markers, and glucose tolerance in genetically obese ob/ob mice 66. The observed reduction in glucose levels and prevention of fat mass accumulation was later attributed to upregulation of heme oxygenase expression and downregulation of endocannabinoid receptor-1 expression resulting in adipose tissue remodeling 67. In addition, the authors detected increased AKT and AMPK phosphorylation in the aortas of L-4F treated mice, which could be prevented by inhibitors of phosphatidylinositol 3

(PI3) kinase activity. However, pharmacological inhibition only partially prevented the glucose lowering effect of L-4F in ob/ob mice, pointing to possible alternative mechanisms. Another recent study in mice showed that apoA-I and its mimetic D-4F, a

13 absorbable L-4F stereoisomer, increased energy expenditure by upregulating the expression of UCP-1 in brown adipose tissue, thus adding an ulterior anti-obesity function to HDL 68.

Anti-apoptotic Functions. HDL has been demonstrated to inhibit apoptotic activity in at least six different cell types including vascular endothelial and smooth muscle cells, some leukocytes, pancreatic beta cells, cardiomyocytes and even bone-forming osteoblasts. Several different modes of action have been identified involving both the protein and lipid components of HDL which can act directly to influence cellular signaling or by various indirect mechanisms to prevent apoptosis. The protein component of HDL, consisting primarily of apoA-I, has been shown to be responsible for about 70% of HDL mediated inhibition of oxLDL induced apoptosis in human microvascular endothelial cells (HMEC-1) 69. There is evidence that some of the less abundant HDL proteins may also be involved. One minor HDL protein called paraoxonase 1 (PON 1) has been found to be important for HDL binding and protection of macrophages from apopotic events. This effect is the result of increased expression of SR-BI via PON 1-mediated activation of ERK1/2 and PI3K signaling pathways 70.

Another HDL associated protein, alpha-1-antitrypsin (A1AT), can prevent apoptosis in a more indirect manner. This protein inhibits elastase-induced degradation of the extracellular matrix, thus preventing detachment and subsequent apoptosis of vascular smooth muscle cells 71. The lipid composition of HDL can also play roles in anti- apoptotic signaling. Recently, reports have indicated important roles for sphingosine 1 phosphate (S1P), a common component of HDL, in the protection of cardiomyocytes,

14 pancreatic beta cells and endothelial cells from apoptosis 72-75. In addition to SR-BI, these protein and lipid components of HDL are likely to interact with other cell surface proteins to produce these effects. For example, the ATP binding cassette transporters

ABCA1 and ABCG1 have both been implicated in HDL’s anti-apoptotic effects on macrophages 76, 77. Furthermore, a known receptor for apoA-I and HDL called cell surface F1-ATPase (an enzyme related to mitochondrial F1 ATPase) can inhibit apoptosis independent of ABC transporters and SR-BI. Interaction of F1-ATPase with

HDL not only protects against apoptotic signaling but also stimulates endothelial cell proliferation 78. Different density fractions of HDL can have varying anti-apoptotic capacities with small dense HDL3c having the most potent effects, shown on endothelial cells and osteoblasts 69, 79. Interestingly, this activity is impaired in patients with metabolic syndrome, possibly due to changes in HDL particle protein and lipid composition seen in these patients 80.

Influence on Stem Cells and Embryogenesis. Another functional property of HDL currently under investigation is a role in the maturation of stem cells. Bone marrow cells

(BMCs) are a key source of vascular progenitor cells that contribute to vessel repair upon endothelial denudation. BMCs are thought to be constantly shuffling back and forth between the bone marrow and the circulation, allowing them to migrate to sites of injury in response to pro-inflammatory cytokines. Once there, they can differentiate into cell types that are needed to effect endothelial repair. It has been observed that treatment of lineage-negative BMCs with lipid-free apoA-I induced a change in their morphology and promoted expression of CD31, an adhesion molecule present in

15 endothelial cells. This increased their ability to bind to both fibronectin and cultured endothelial cells 81. A mutant of apoA-I lacking a key lipid binding site failed to promote these transformations, suggesting that apoA-I may mediate these effects via lipid efflux.

The authors suggested that apoA-I stimulation of BMC differentiation may be a mechanism by which HDL mediates vessel repair. The ability of apoA-I to mobilize cellular lipids also appears to play a role in the proliferation of hematopoietic stem and progenitor cells (HSPCs). For example, it was found that ABCA1 and ABCG1 deficient

(ABCA1-/- ABCG1-/-) mice, which have extremely low levels of circulating HDL, displayed a five-fold increase in hematopoietic stem and progenitor cells (HSPCs) compared to controls 82. These levels decreased significantly when ABCA1-/- ABCG1-/- bone marrow was transplanted into apoA-I transgenic mice, which display elevated levels of HDL.

These results suggest that HDL may inhibit leukocytosis (elevated circulating leukocytes) and reduce monocyte infiltration into vascular lesions.

Additionally, HDL is the predominant lipoprotein in follicular fluid and may play a role in oocyte development and embryogenesis (reviewed in 83). For example, a negative correlation has been found between human follicular HDL-cholesterol levels and embryo fragmentation during in vitro fertilization 84. HDL metabolism may also affect embryo viability in vivo. Liver specific SR-BI knockout female mice with significantly decreased HDL uptake by the liver and large dysfunctional circulating HDL were infertile. However, reconstitution of SR-BI expression in the knockout animals by adenovirus-mediated gene delivery restored both HDL morphology and fertility 85, consistent with a role for HDL in female reproductive physiology.

Effects on Platelet Function. HDL has been known to affect platelet function for many years, though the effects have been complex and variable between experimental systems (for a recent review, see 86). Recent evidence has revealed that the cholesterol homeostatic functions of HDL and apoA-I may significantly impact platelet function. Mice that lack SR-BI are known to be thrombocytopenic (low platelet count).

Their platelets exhibit reduced ability to aggregate and this has been correlated to abnormally high levels of free cholesterol in the cells 87. This suggests that platelet SR-

BI interactions with HDL can modulate cellular cholesterol levels for optimal platelet function. This assertion was supported by a more recent study in which T2D patients were infused with reconstituted HDL preparations 88. The infused patients exhibited a

50% decrease in platelet aggregation response vs. controls. The effect was attributed to the phospholipid fraction of the particles with apoA-I having no role, consistent with the idea that SR-BI-mediated cholesterol efflux can beneficially modulate platelet function. The discovery of several HDL proteins with known roles in platelet function such as the complement family and platelet activating factor acylhydrolase, strongly suggest that more ties between HDL and platelet function exist which go beyond cholesterol efflux. More work in this area is clearly needed.

Current measures of HDL subfractions

In the basic science laboratory and the clinic, HDL is referred to by a complex and sometimes confusing set of definitions. Numerous modes of HDL separation from plasma have led to separate nomenclature systems that center on a specific physico-

17 chemical or immunological property. For example, HDL has been classically defined by its hydrated density (driven by the ratio of protein to lipid in the particles) as separated by ultracentrifugation. In humans, it is typically found as two major forms: HDL2 (d =

1.063-1.125 g/ml) and HDL3 (d= 1.125-1.210 g/ml) with diameters ranging from 70-120

Å. These have been further refined to five subspecies HDL2(b, a) and HDL3(a, b, c) using tighter density cuts 89. Given the method’s ease and wide use, the density-based nomenclature has been the most generally adopted. However, HDL has also been classified by major apolipoprotein content using immunoaffinity chromatography into apoA-I containing particles that lack apoA-II (LpA-I) and those that contain both apoA-I and apoA-II (LpA-I/A-II) 90. By contrast, agarose gel electrophoresis can classify HDL by charge density into alpha, pre-alpha or pre-beta forms depending on the degree of negative charge 91. In the clinic, human plasma is commonly treated with phosphotungstic acid or heparin and magnesium chloride to precipitate the apoB- containing lipoproteins LDL and VLDL 92. The precipitated lipoproteins are pelleted and cholesterol in the supernate is referred to as HDL cholesterol (HDL-C). It is this HDL-C measurement that underlies the majority of epidemiological studies illustrating the inverse association with CAD. Other physical properties such as molecular size (gel filtration chromatography) 93 and ionic character (ion exchange chromatography) 94 have been exploited for lipoprotein lipid analyses, but have not yet been widely used for proteomics due to HDL coelution with unrelated high abundance plasma proteins.

Attempts to sub-classify HDL have been driven by the promise that a better understanding of HDL subspeciation will yield a more detailed knowledge of its metabolism and perhaps predictive biomarkers for CAD. However, the functional basis

18 for the high compositional heterogeneity of HDL has largely remained a mystery. One reason for this is that there is little concordance between the different HDL separation methodologies. For example, a sample of LpA-I as isolated by immunoaffinity chromatography contains particles of HDL2 and HDL3 density as well as alpha and pre- beta electrophoretic species and vice versa. Thus, functional characterizations of HDL subspecies tend to be framed only within the context of the method used to isolate them. An HDL subfraction isolated by a particular method can be considered a true subfraction only in the physico-chemical sense, not necessarily the functional sense.

Another hurdle has been that functional conclusions about a given HDL subfraction can be compromised by any artifacts inherent to the method used to isolate it. It is increasingly clear that the high shear forces and ionic strengths experienced during KBr density ultracentrifugation strip off or redistribute certain proteins. For example, van’t

Hooft et al. showed that ultracentrifugally isolated HDL interacted more avidly to the liver apoE receptor than gel filtered HDL because of altered apoE content or depletion of other apolipoproteins 95. Another study showed that the use of deuterium oxide and sucrose in ultracentrifugation separations resulted in an altered HDL proteome compared to the classic salt-based method 96. Thus, KBr centrifugation may favor ‘core proteins’ that survive the spins, but more transiently associated proteins – which may serve important biological functions – may be underestimated, found in nonphysiological locations or missed completely. Immunoaffinity isolation of lipoproteins holds the promise of more gentle separation conditions than ultracentrifugation. However, this approach has its own set of problems, chiefly the bias introduced by the specificity of the antibodies used in the separation. Given its abundance, many have been tempted

19 to define HDL by the presence of apoA-I. Unfortunately, apoA-I is also present on LDL- and VLDL-sized particles 97 and lipidated particles that cofractionate with HDL have been shown to lack both apoA-I and apoA-II 98. Thus, immunoprecipitation techniques that target a single protein may not cast a wide enough net for capturing all HDL particles. Finally, despite being epidemiologically predictive, the clinical HDL-C measurement fails to take into account the protein complement or any particle subspeciation within HDL. Indeed, it is not clear exactly which HDL subspecies are captured by the precipitation techniques. For these reasons, prominent clinicians such as Dan Rader have argued that the HDL-C paradigm is “insufficient to capture the functional variation in HDL particles” 99.

Applications of modern proteomics to HDL.

Since HDL was first isolated by ultracentrifugation over 60 years ago, hundreds of biochemical and immunological studies have been directed at characterizing its protein constituents. In addition to structural stability, these proteins (called apolipoproteins) impart biological directionality to the lipid cargo by a) targeting it to various tissues, b) modifying its chemical form (i.e. lipolysis or esterification), or c) transferring it to other lipoproteins. Roughly 65% of HDL protein mass is comprised of apolipoprotein (apo) A-

I with another 15% by apoA-II. The remainder includes around 50 proteins that are each too low in abundance to be present on all circulating HDL particles. The potential for differential accumulation of these proteins into certain subspecies likely drives the well-known polydispersity of HDL.

Prior to 2005, HDL was generally thought to contain proteins in four major functional groups: 1) proteins associated with lipid transport or lipoprotein integrity, i.e. the ‘apos”, apoA-I, apoA-II, apoC’s, apoD, apoF, apoM, apoE, etc., 2) lipolytic enzymes such as LCAT, paraoxonase, platelet activating factor aryl hydrolase, etc., 3) Lipid transfer proteins including cholesteryl ester transfer protein (CETP) and phospholipid transfer protein, and 4) acute phase response proteins such as serum amyloid A, clusterin, and apoA-IV. The presence of these proteins fit well into the general dogma of a primary role for HDL in RCT.

However, this view has undergone significant revision in light of recent proteomics studies on HDL. With the advent of soft ionization techniques and the resulting leaps in mass spectrometry (MS) technology, modern proteomic approaches have become powerful tools for mapping complex protein mixtures. The advantage of proteomics over standard biochemical approaches lies in the ability to identify proteins without prior suspicion of their presence in HDL. Though strategies are almost as diverse as the laboratories performing them, the proteomic approaches applied to HDL can be classified into two general categories. The first uses gel electrophoresis to spread HDL proteins either by size only (1 Dimensional, 1D) or by charge and then size

(2 Dimensional, 2D). The resulting gel spots are excised, digested with a protease such as trypsin and identified by high resolution tandem MS. In the second approach, sometimes referred to as a ‘shotgun’ technique, tryptic peptides are first generated from

HDL in solution. The peptides are then separated by HPLC (commonly by reverse phase, but multidimensional separations are also used), subjected to electrospray ionization (ESI) as they elute from the column and analyzed by tandem MS. Note: a

21 more detailed description of proteomic technologies can be found here 100. To date, there have been nine major applications of modern proteomic approaches to human

HDL. The findings are compiled in Table 1-1.

Using a 2D electrophoresis approach (isoelectric focusing followed by SDS-

101 PAGE), Karlsson et al identified 13 proteins in pools of HDL2 and HDL3 separated by two-step discontinuous density ultracentrifugation from a pool of healthy donors. Of these, 11 were known HDL components by biochemical methods. One of the novel findings was α-1-antitrypsin (A1AT), a serine protease inhibitor (serpin) since confirmed by others. This protein is known to modulate activated neutrophils and thus supported a possible role for HDL in innate immunity 102. In addition, this study highlighted a key advantage of the charge dimension of the electrophoretic approach in that multiple isoforms of apoA-I, apoA-II, apoC-III, apoE, apoM, SAA and SAA-4 were identified that varied with respect to the presence of a propeptide, potential glycosylation, or sialyation.

Another study in the same year utilized both the shotgun approach as well as a different

2D electrophoresis strategy (native separation followed by SDS-PAGE) to study density

103 isolated fractions roughly corresponding to HDL2 and HDL3 . The study yielded 24 proteins including most found by Karlsson et al., as well as A1AT. Serotransferrin was postulated as a new HDL-associated protein and alpha-2-macroglobulin, another protease inhibitor, was found in confirmation of previous biochemical reports.

Rezaee et al. 104 used a multipronged approach of 1D and 2D electrophoresis, shotgun proteomic methods and immunological assays to study centrifugally isolated total HDL from normal donors. This was the first study to employ an isotope-coded affinity tag (ICAT) method to identify lower abundance proteins that may be masked by

22 the more common constituents. In this strategy, tryptic peptides containing Cys residues were derivatized with a heavy or light isotopic tag. By selecting peptides with a threshold ratio of light to heavy label, the authors identified up to 40 Cys-containing HDL proteins. While many of these have yet to be confirmed, this study first identified key mediators of the complement system, C3, C1 inhibitor and complement factor H, implicating HDL in hemostasis and innate immunity. That same year, Hortin et al. 105 examined ultracentrifugally isolated HDL from normal donors, but focused on low molecular weight peptides <4000 Da - finding many of the same proteins (or at least pieces of them). The finding of small peptides from fibrinogen suggested HDL may be a circulating reservoir for proteolytic fragments of major plasma proteins. This study pointed out a potential pitfall of bottom-up (i.e. peptide based) MS methods in that peptide sequence identifications do not necessarily indicate that the entire protein is present in HDL.

The most comprehensive proteomic analysis of HDL to date was performed by

106 Vaisar et al . Using centrifugally isolated total HDL or HDL3 from normal subjects, they used multidimensional protein identification technology (MUDPIT) and ESI-MS to identify 48 HDL proteins, 35 previously known and 13 new identifications. Four of the new HDL associations play major roles in complement regulation, strengthening the argument for a role of HDL in innate immunity. Others showed a clear theme of protease inhibition. Most importantly, this study took HDL proteomics to the next level by comparing protein profiles of six normolipidemic control subjects to seven subjects with established CAD. Using a novel peptide counting strategy, it was found that HDL3 from CAD subjects was enriched in apoC-IV, PON1, complement C3, apoA-IV and

23 apoE. A follow-up study compared the HDL3 proteome of subjects with stenotic lesions verified by angiography before and one year after combination treatment with a statin and the HDL raising drug niacin 107. The treatment was found to reduce apoE levels while increasing apoF and PLTP, partially remodeling the stenotic proteomic profile toward that of control subjects. A later study by Alwaili et al used shotgun proteomics to compare the HDL proteome between patients with stable CAD versus those with acute coronary syndrome (ACS) 108. This study identified 67 HDL associated proteins of which 5 were novel and showed that the HDL proteome shifts toward a more pro- inflammatory profile in ACS having higher levels of SAA and complement C3 among other inflammatory proteins. One additional study comparing HDL proteomes between disease states has been done. Watanabe et al used immunoaffinity capture to isolate

HDL from rheumatoid arthritis patients with either anti or proinflammatory HDL based on the inflammatory index, a measure based on anti-oxidative capacity 109. Similar to the previously mentioned study, here they found elevated levels of SAA but also three other proteins apoJ, fibrinogen and haptoglobin were increased in RA patients determined to have pro-inflammatory HDL. The findings of these studies comparing HDL associated proteins in different disease states suggest that different types of conditions in the body can alter the HDL proteome differently.

Evidence for HDL subspeciation based on protein content.

The proteins that comprise PL-rich LPs are called exchangeable apolipoproteins due to their ability to associate with and dissociate from lipoproteins as a function of

24 their relative affinities for the particles vs. a stable lipid-free form. Indeed, careful kinetic studies have documented the movement of apolipoproteins between VLDL and HDL 110.

Given this reversibility, one might be tempted to think of HDL as a transient ensemble of proteins randomly exchanging between lipid assemblies. Indeed, certain transfer proteins and enzymes likely ping-pong between different particles to perform their functions. Indeed, the recent X-ray crystal structure of CETP indicates that the boomerang-shaped protein acts as a mobile conduit that can move CE and TG between

HDL and TG-rich LPs 111. However, there is significant evidence that many apolipoproteins do in fact segregate into compositionally stable particles. Asztalos and colleagues have made extensive use of a native 2D gel electrophoresis system that separates human plasma first by charge and then by size. The gels are then probed with antibodies to visualize protein migration patterns. ApoA-I appears in up to 11 distinct spots that represent highly negatively charged species of various size (pre-α1,

2, 3), moderately negatively charged (α1, 2, 3) and less negative charged species

(preβ1 a,b and preβ2, a,b,c), all of various diameters 112. When probing for other HDL- associated proteins, highly distinct patterns emerged 98. ApoA-II associated with apoA-I in the α2 and α3 species, but not in the others. ApoE showed up on larger particles that also failed to overlap completely with apoA-I. They also found evidence for apoA-IV- only lipoproteins of similar diameter to apoA-I-containing particles. Furthermore, our proteomic characterization of PL-rich LP density subspecies also showed clearly distinct abundance patterns for the different proteins 113. Some preferred small, dense particles while others preferred large, light ones. These reports indicate that apolipoproteins are not randomly exchanging in plasma. The mechanisms driving such segregation are

25 unknown but could be related to a) affinity of a given protein to a given lipid composition or degree of surface curvature, or b) specific protein to protein interactions on the particle surface that maintain protein segregation.

This latter idea of specific protein-protein interactions in PL-rich LPs opens the possibility for cooperative function. There is ample evidence that major activities of HDL rely on cooperative interactions between proteins cohabitating on a particle. The classic example is the relationship of apoA-I and LCAT. On its own, LCAT is relatively inefficient in mediating cholesterol esterification in lipoproteins. However, apoA-I acts as a cofactor to stimulate this activity by several orders of magnitude when present on the same particle 114. ApoF, also known as lipid transport inhibitor protein, can inhibit the CETP-mediated exchange of CE between HDL and TG-rich LPs, possibly by modulating CETP’s affinity for the HDL particle surface 115. There is also growing evidence that one role of apoA-II may be to modulate apoA-I function. A study compared the hydrolysis rates of LpA-I, LpA-II and LpA-I/A-II HDL by endothelial lipase

(EL), an important enzyme for the physiological regulation of HDL-C levels 116. The lipid hydrolysis rate was highest in the mixed particles but lower in LpA-I and almost undetectable in LpA-II. That apoA-II facilitated lipid hydrolysis in mixed particles but not in LpA-II suggests it can affect apoA-I conformation to modulate EL activation.

However, the most striking example of on-particle protein cooperation relates to HDL’s role in innate immunity. It was recognized in 1978 that a dense fraction of HDL could mediate the lysis of T. brucei, a trypanosome that causes African sleeping sickness.

This activity was referred to as trypanosome lytic factor (TLF) 117. Immunoprecipitation studies demonstrated that TLF is a specific HDL subparticle that contains apoA-I, apoL-I

26 and haptoglobin-related protein (HRP). The current model for TLF lysis of T. brucei holds that the HDL particle is taken up by the trypanosome in a receptor-mediated pathway, possibly via the HRP moiety (for a review, see 118). The complex is then targeted to the lysosome where apoL-I, via a colicin-like pore forming domain, permeates the organelle to kill the organism. Interestingly, apoA-I may be required for the proper sequestration of apoL-I and HRP to form TLF. TLF’s unique composition is the strongest evidence yet that distinct particles within classically defined HDL exist and perform highly specialized functions that are quite distinct from traditional lipid transport roles 119. Given the large number of proteins in play, it is easy to imagine many more

PL-rich LPs yet to be discovered.

More recently, we applied a shotgun ESI-MS/MS approach to learn about the distribution of proteins across five different density sub-fractions (HDL2b,a and

HDL3a,b,c) isolated by density ultracentrifugation 113. Using an abundance pattern analysis, we categorized 28 different proteins into five groups based on their distribution across the fractions. A correlational network suggested the existence of distinct protein clusters that may be suggestive of PL-rich LP subpopulations defined by specific protein interactions. Also, levels of apoL-I and paraoxonase were correlated with the capacity of HDL to protect against LDL oxidation. Taking subfraction analysis to the next level we used the same proteomics approach to analyze 17 lipoprotein subfractions isolated by size using gel filtration chromatography 120 (data presented in Chapter 2).

HDL subspeciation and its impact on the treatment of cardiovascular disease.

Statin therapies, though successful commercially, only reduce CAD risk by about

1/3 121. Many have pointed to this as evidence for the need to raise HDL-C in combination with the statin-mediated lowering of LDL. Unfortunately, HDL is sometimes thought of in rather simplistic terms. While its compositional heterogeneity is well cited, it tends to be discussed as a single entity (i.e. HDL-C), most often tied to reverse cholesterol transport. Indeed, current pharmacological therapies such as niacin or those in development including CETP inhibition and apoA-I transcription stimulation aim to raise plasma HDL cholesterol in the generic sense without direct knowledge of the functionality (or lack thereof) of the elevated species. While there is no question that high plasma levels of HDL-C are inversely correlated with CVD on a population basis, there are also many individuals with high HDL-C who have CVD and vice versa. The implication of this is clear - not all HDL is created equal. Among the numerous particle populations comprising HDL, some may be cardioprotective, some may not 122. It is quite possible that raising HDL-C indiscriminately may increase the wrong particles at the expense of cardioprotective ones. We need a better understanding of the subparticle makeup of the fractions classically referred to as “HDL”. With such knowledge, perhaps more targeted therapies can be brought online to raise plasma levels of the cardioprotective PL-rich LPs. It is possible, maybe even likely, that a successful therapeutic strategy may not even raise HDL-C as it is measured in the clinic currently. Alternatively, small molecule therapies could be explored that mimic the cardioprotective effects of identified beneficial PL-rich subspecies. Given the tremendous resources that have been poured into non-specific raising of plasma HDL-

C, and the relatively meager successes to date, it seems prudent to invest more research into understanding HDL subspeciation.

The application of high resolution MS-based proteomics has dramatically increased our understanding of HDL protein complexity. These findings leave little doubt that HDL is not only involved in lipid transport, but also proteinase inhibition, anti- inflammation, complement regulation and innate immunity. Comparative proteomics studies offer the hope that new biomarkers can be identified to predict not only CAD but other inflammatory diseases as well. In addition to monitoring the ups and downs of particular proteins, we may find that the overall levels of a given protein may not differ between healthy and diseased subjects, but that its distribution among PL-rich LPs or its associations with other proteins may change significantly. Realization of these goals will require solving technical problems including increasing sample throughput to boost the number of subjects that can be efficiently studied as well as the development of more sensitive and quantitative comparison strategies. Another challenge relates to the fact that, prior to this thesis, all detailed proteomic characterizations had been performed on HDL isolated by ultracentrifugation. Given the potential for artifacts, these studies must be verified using alternative separations. Unfortunately, methods such as gel filtration, ion exchange or native gel electrophoresis are compromised by the presence of high abundance contaminants that are not at issue in ultracentrifugation.

New technologies for either lipoprotein isolation or improved MS dynamic range will be required to meet this important challenge.

The global hypothesis driving this thesis is that: The total pool of HDL in an individual is composed of numerous proteomically distinct subspecies and that their compositional diversity will result in varying functional capacities. The goals of this thesis are to develop new methods to isolate and analyze these subspecies so that we can determine their compositions and functional roles in health and in a specific metabolic disease, type 2 diabetes. By understanding how individual subspecies are composed and function in health and in sickness we can better utilize them as biomarkers for disease risk or target them specifically for therapeutic benefit.

Goals of thesis research

Aim 1: To identify and characterize HDL subspecies based on protein composition.

Hypothesis: The pool of total HDL in an individual is composed of several stable and proteomically distinct HDL subspecies.

Using three different biochemical separation techniques, each exploiting a different physical property, fresh human plasma will be fractionated. The protein distribution across HDL containing fractions will be determined by electrospray ionization tandem mass spectrometry (ESI-MS/MS) and distribution profiles produced by each isolation technique will be analyzed by a novel correlation strategy to identify protein groupings that persist across the different methods. Candidate protein interactions identified by this correlation analysis will be confirmed by chemical crosslinking studies.

Aim 2: Analyze functional heterogeneity across separated plasma HDL fractions.

Hypothesis: Variation in the proteomic composition of HDL particles results in different functional capacities of each subspecies.

We will use a panel of functional assays to measure the level of activity of known HDL functions in collected HDL sub-fractions. The functions to be examined are: a) ability to prevent oxidation of low density lipoprotein (LDL) particles, b) ability to promote cholesterol efflux from macrophages and c) the ability to suppress vascular inflammation. Functional profiles will be analyzed using the correlation strategy used in

Aim 1 to correlate specific functions with individual or pairs of HDL proteins. In the long term, reconstituted particles mimicking these subspecies can be tested for function in vivo in genetically manipulated mouse models.

Aim 3: Examine the effect of type 2 diabetes on HDL subspecies distribution in young adults.

Hypothesis: HDL composition is altered in patients with type 2 diabetes resulting in decreased protection from cardiovascular disease.

Plasma collected from healthy or type 2 diabetic adolescent and young adult males will be separated by gel filtration chromatography to fractionate lipoprotein subpopulations.

Collected fractions will be analyzed for lipid composition using enzymatic assays and for protein composition using mass spectrometry. Compositional data from healthy

31 participants will be compared to that from type 2 diabetics. Additionally, we will examine correlations between HDL composition and measures of arterial stiffness and thickness taken at the time of HDL collection.

Anticipated effect of research on lipoprotein field

This work will result in the following advances: 1) We will develop new plasma separation techniques optimized for the isolation of phospholipid rich lipoproteins that are compatible with high resolution mass spectrometry to more fully understand the complexity of the HDL proteome. Current proteomic studies have been restricted solely to ultracentrifugally isolated HDL. 2) We will identify for the first time new subspecies of

HDL using a completely unbiased and novel approach. 3) We will be able to link these subspecies with specific functions of HDL. 4) We will be able to link alterations in HDL subspecies distribution to a specific disease state and correlate these changes with direct clinical measures of cardiovascular health.

Reference List

1 BARR DP, RUSS EM, EDER HA. Protein-lipid relationships in human plasma. II. In

atherosclerosis and related conditions. Am J Med 1951;11(4):480-93.

2 Gordon T, Castelli WP, Hjortland MC, Kannel WB, Dawber TR. High density

lipoprotein as a protective factor against coronary heart disease. The Framingham

Study. Am J Med 1977;62(5):707-14.

3 Gordon DJ, Probstfield JL, Garrison RJ, Neaton JD, Castelli WP, Knoke JD,

Jacobs DR, Jr., Bangdiwala S, Tyroler HA. High-density lipoprotein cholesterol and

cardiovascular disease. Four prospective American studies. Circulation

1989;79(1):8-15.

4 Chirovsky DR, Fedirko V, Cui Y, Sazonov V, Barter P. Prospective studies on the

relationship between high-density lipoprotein cholesterol and cardiovascular risk: a

systematic review. Eur J Cardiovasc Prev Rehabil 2009;16(4):404-23.

5 Fan AZ, Dwyer JH. Sex differences in the relation of HDL cholesterol to

progression of carotid intima-media thickness: the Los Angeles Atherosclerosis

Study. Atherosclerosis 2007;195(1):e191-e196.

6 Upmeier E, Lavonius S, Lehtonen A, Viitanen M, Isoaho H, Arve S. Serum lipids

and their association with mortality in the elderly: a prospective cohort study. Aging

Clin Exp Res 2009;21(6):424-30.

7 Goldbourt U, Yaari S, Medalie JH. Isolated low HDL cholesterol as a risk factor for

coronary heart disease mortality. A 21-year follow-up of 8000 men. Arterioscler

Thromb Vasc Biol 1997;17(1):107-13.

8 Okamura T, Hayakawa T, Kadowaki T, Kita Y, Okayama A, Ueshima H. The

inverse relationship between serum high-density lipoprotein cholesterol level and

all-cause mortality in a 9.6-year follow-up study in the Japanese general

population. Atherosclerosis 2006;184(1):143-50.

9 Coca A, Cea-Calvo L, Lozano JV, Inaraja V, Fernandez-Perez C, Navarro J, Bonet

A, Redon J. High-density lipoprotein cholesterol and cardiovascular disease in

Spanish hypertensive women. The RIMHA study. Rev Esp Cardiol

2009;62(9):1022-31.

10 Hayashi T, Kawashima S, Itoh H, Yamada N, Sone H, Watanabe H, Hattori Y,

Ohrui T, Yokote K, Nomura H, Umegaki H, Iguchi A. Low HDL cholesterol is

associated with the risk of stroke in elderly diabetic individuals: changes in the risk

for atherosclerotic diseases at various ages. Diabetes Care 2009;32(7):1221-3.

11 Topakian R, Sonnberger M, Nussbaumer K, Haring HP, Trenkler J, Aichner FT.

Postprocedural high-density lipoprotein cholesterol predicts carotid stent patency

at 1 year. Eur J Neurol 2008;15(2):179-84.

12 Cotter AG, Satchell CS, O'Halloran JA, Feeney ER, Sabin CA, Mallon PW. High-

density lipoprotein levels and 10-year cardiovascular risk in HIV-1-infected

patients. AIDS 2011;25(6):867-9.

13 Menotti A, Lanti M, Zanchetti A, Botta G, Laurenzi M, Terradura-Vagnarelli O,

Mancini M. The role of HDL cholesterol in metabolic syndrome predicting

cardiovascular events. The Gubbio population study. Nutr Metab Cardiovasc Dis

2011;21(5):315-22.

14 Jafri H, Alsheikh-Ali AA, Karas RH. Baseline and on-treatment high-density

lipoprotein cholesterol and the risk of cancer in randomized controlled trials of lipid-

altering therapy. J Am Coll Cardiol 2010;55(25):2846-54.

15 Singh-Manoux A, Gimeno D, Kivimaki M, Brunner E, Marmot MG. Low HDL

cholesterol is a risk factor for deficit and decline in memory in midlife: the Whitehall

II study. Arterioscler Thromb Vasc Biol 2008;28(8):1556-62.

16 Hooper AJ, Burnett JR. Dalcetrapib , a cholesteryl ester transfer protein modulator.

Expert Opin Investig Drugs 2012;21(9):1427-32.

17 Duggal JK, Singh M, Attri N, Singh PP, Ahmed N, Pahwa S, Molnar J, Singh S,

Khosla S, Arora R. Effect of niacin therapy on cardiovascular outcomes in patients

with coronary artery disease. J Cardiovasc Pharmacol Ther 2010;15(2):158-66.

18 Voight BF, Peloso GM, Orho-Melander M, Frikke-Schmidt R, Barbalic M, Jensen

MK, Hindy G, Holm H, Ding EL, Johnson T, Schunkert H, Samani NJ, Clarke R,

Hopewell JC, Thompson JF, Li M, Thorleifsson G, Newton-Cheh C, Musunuru K,

Pirruccello JP, Saleheen D, Chen L, Stewart A, Schillert A, Thorsteinsdottir U,

Thorgeirsson G, Anand S, Engert JC, Morgan T, Spertus J, Stoll M, Berger K,

Martinelli N, Girelli D, McKeown PP, Patterson CC, Epstein SE, Devaney J,

Burnett MS, Mooser V, Ripatti S, Surakka I, Nieminen MS, Sinisalo J, Lokki ML,

Perola M, Havulinna A, de FU, Gigante B, Ingelsson E, Zeller T, Wild P, de Bakker

PI, Klungel OH, Maitland-van der Zee AH, Peters BJ, de BA, Grobbee DE,

Kamphuisen PW, Deneer VH, Elbers CC, Onland-Moret NC, Hofker MH,

Wijmenga C, Verschuren WM, Boer JM, van der Schouw YT, Rasheed A, Frossard

P, Demissie S, Willer C, Do R, Ordovas JM, Abecasis GR, Boehnke M, Mohlke KL,

Daly MJ, Guiducci C, Burtt NP, Surti A, Gonzalez E, Purcell S, Gabriel S, Marrugat

J, Peden J, Erdmann J, Diemert P, Willenborg C, Konig IR, Fischer M,

Hengstenberg C, Ziegler A, Buysschaert I, Lambrechts D, Van de Werf F, Fox KA,

El Mokhtari NE, Rubin D, Schrezenmeir J, Schreiber S, Schafer A, Danesh J,

Blankenberg S, Roberts R, McPherson R, Watkins H, Hall AS, Overvad K, Rimm

E, Boerwinkle E, Tybjaerg-Hansen A, Cupples LA, Reilly MP, Melander O,

Mannucci PM, Ardissino D, Siscovick D, Elosua R, Stefansson K, O'Donnell CJ,

Salomaa V, Rader DJ, Peltonen L, Schwartz SM, Altshuler D, Kathiresan S.

Plasma HDL cholesterol and risk of myocardial infarction: a mendelian

randomisation study. Lancet 2012;380(9841):572-80.

19 Lund-Katz S, Phillips MC. High density lipoprotein structure-function and role in

reverse cholesterol transport. Subcell Biochem 2010;51:183-227.

20 Zhang Y, Zanotti I, Reilly MP, Glick JM, Rothblat GH, Rader DJ. Overexpression of

apolipoprotein A-I promotes reverse transport of cholesterol from macrophages to

feces in vivo. Circulation 2003;108(6):661-3.

21 Cockerill GW, Rye KA, Gamble JR, Vadas MA, Barter PJ. High-density lipoproteins

inhibit cytokine-induced expression of endothelial cell adhesion molecules.

Arteriosclerosis, Thrombosis & Vascular Biology 1995;15(11):1987-94.

22 Patel S, Di Bartolo BA, Nakhla S, Heather AK, Mitchell TW, Jessup W, Celermajer

DS, Barter PJ, Rye KA. Anti-inflammatory effects of apolipoprotein A-I in the rabbit.

Atherosclerosis 2010.

23 Nobecourt E, Tabet F, Lambert G, Puranik R, Bao S, Yan L, Davies MJ, Brown BE,

Jenkins AJ, Dusting GJ, Bonnet DJ, Curtiss LK, Barter PJ, Rye KA. Nonenzymatic

glycation impairs the antiinflammatory properties of apolipoprotein A-I. Arterioscler

Thromb Vasc Biol 2010;30(4):766-72.

24 Tolle M, Pawlak A, Schuchardt M, Kawamura A, Tietge UJ, Lorkowski S, Keul P,

Assmann G, Chun J, Levkau B, van der Giet M, Nofer JR. HDL-associated

lysosphingolipids inhibit NAD(P)H oxidase-dependent monocyte chemoattractant

protein-1 production. Arterioscler Thromb Vasc Biol 2008;28(8):1542-8.

25 Mineo C, Deguchi H, Griffin JH, Shaul PW. Endothelial and antithrombotic actions

of HDL. Circ Res 2006;98(11):1352-64.

26 Drew BG, Fidge NH, Gallon-Beaumier G, Kemp BE, Kingwell BA. High-density

lipoprotein and apolipoprotein AI increase endothelial NO synthase activity by

protein association and multisite phosphorylation. Proc Natl Acad Sci U S A

2004;101(18):6999-7004.

27 Mineo C, Yuhanna IS, Quon MJ, Shaul PW. High density lipoprotein-induced

endothelial nitric-oxide synthase activation is mediated by Akt and MAP kinases. J

Biol Chem 2003;278(11):9142-9.

28 Norata GD, Callegari E, Marchesi M, Chiesa G, Eriksson P, Catapano AL. High-

density lipoproteins induce transforming growth factor-beta2 expression in

endothelial cells. Circulation 2005;111(21):2805-11.

29 Yuhanna IS, Zhu Y, Cox BE, Hahner LD, Osborne-Lawrence S, Lu P, Marcel YL,

Anderson RG, Mendelsohn ME, Hobbs HH, Shaul PW. High-density lipoprotein

binding to scavenger receptor-BI activates endothelial nitric oxide synthase. Nat

Med 2001;7(7):853-7.

30 Tarr PT, Tarling EJ, Bojanic DD, Edwards PA, Baldan A. Emerging new paradigms

for ABCG transporters. Biochim Biophys Acta 2009;1791(7):584-93.

31 Terasaka N, Westerterp M, Koetsveld J, Fernandez-Hernando C, Yvan-Charvet L,

Wang N, Sessa WC, Tall AR. ATP-Binding Cassette Transporter G1 and High-

Density Lipoprotein Promote Endothelial NO Synthesis Through a Decrease in the

Interaction of Caveolin-1 and Endothelial NO Synthase. Arterioscler Thromb Vasc

Biol 2010;30(11):2219-25.

32 Whetzel AM, Sturek JM, Nagelin MH, Bolick DT, Gebre AK, Parks JS, Bruce AC,

Skaflen MD, Hedrick CC. ABCG1 deficiency in mice promotes endothelial

activation and monocyte-endothelial interactions. Arterioscler Thromb Vasc Biol

2010;30(4):809-17.

33 Murphy AJ, Woollard KJ, Hoang A, Mukhamedova N, Stirzaker RA, McCormick

SP, Remaley AT, Sviridov D, Chin-Dusting J. High-density lipoprotein reduces the

human monocyte inflammatory response. Arterioscler Thromb Vasc Biol

2008;28(11):2071-7.

34 Tang C, Vaughan AM, Oram JF. Janus Kinase 2 Modulates the Apolipoprotein

Interactions with ABCA1 Required for Removing Cellular Cholesterol. J Biol Chem

2004;279(9):7622-8.

35 Tang C, Liu Y, Kessler PS, Vaughan AM, Oram JF. The macrophage cholesterol

exporter ABCA1 functions as an anti-inflammatory receptor. J Biol Chem

2009;284(47):32336-43.

36 Okura H, Yamashita S, Ohama T, Saga A, Yamamoto-Kakuta A, Hamada Y,

Sougawa N, Ohyama R, Sawa Y, Matsuyama A. HDL/apolipoprotein A-I binds to

macrophage-derived progranulin and suppresses its conversion into

proinflammatory granulins. J Atheroscler Thromb 2010;17(6):568-77.

37 Murch O, Collin M, Hinds CJ, Thiemermann C. Lipoproteins in inflammation and

sepsis. I. Basic science. Intensive Care Med 2007;33(1):13-24.

38 Jahangiri A. High-density lipoprotein and the acute phase response. Curr Opin

Endocrinol Diabetes Obes 2010;17(2):156-60.

39 Villarroel F, Bastias A, Casado A, Amthauer R, Concha MI. Apolipoprotein A-I, an

antimicrobial protein in Oncorhynchus mykiss: evaluation of its expression in

primary defence barriers and plasma levels in sick and healthy fish. Fish Shellfish

Immunol 2007;23(1):197-209.

40 Caipang CM, Hynes N, Puangkaew J, Brinchmann MF, Kiron V. Intraperitoneal

vaccination of Atlantic cod, Gadus morhua with heat-killed Listonella anguillarum

enhances serum antibacterial activity and expression of immune response genes.

Fish Shellfish Immunol 2008;24(3):314-22.

41 Johnston LD, Brown G, Gauthier D, Reece K, Kator H, Van VP. Apolipoprotein A-I

from striped bass (Morone saxatilis) demonstrates antibacterial activity in vitro.

Comp Biochem Physiol B Biochem Mol Biol 2008;151(2):167-75.

42 Figueiredo PM, Catani CF, Yano T. Serum high-density lipoprotein (HDL) inhibits

in vitro enterohemolysin (EHly) activity produced by enteropathogenic Escherichia

coli. FEMS Immunol Med Microbiol 2003;38(1):53-7.

43 Grunfeld C, Marshall M, Shigenaga JK, Moser AH, Tobias P, Feingold KR.

Lipoproteins inhibit macrophage activation by lipoteichoic acid. J Lipid Res

1999;40(2):245-52.

44 Ma J, Liao XL, Lou B, Wu MP. Role of apolipoprotein A-I in protecting against

endotoxin toxicity. Acta Biochim Biophys Sin (Shanghai) 2004;36(6):419-24.

45 Parker TS, Levine DM, Chang JC, Laxer J, Coffin CC, Rubin AL. Reconstituted

high-density lipoprotein neutralizes gram-negative bacterial lipopolysaccharides in

human whole blood. Infect Immun 1995;63(1):253-8.

46 Jiao YL, Wu MP. Apolipoprotein A-I diminishes acute lung injury and sepsis in mice

induced by lipoteichoic acid. Cytokine 2008;43(1):83-7.

47 Li Y, Dong JB, Wu MP. Human ApoA-I overexpression diminishes LPS-induced

systemic inflammation and multiple organ damage in mice. Eur J Pharmacol

2008;590(1-3):417-22.

48 Singh IP, Chopra AK, Coppenhaver DH, Ananatharamaiah GM, Baron S.

Lipoproteins account for part of the broad non-specific antiviral activity of human

serum. Antiviral Res 1999;42(3):211-8.

49 Srinivas RV, Venkatachalapathi YV, Rui Z, Owens RJ, Gupta KB, Srinivas SK,

Anantharamaiah GM, Segrest JP, Compans RW. Inhibition of virus-induced cell

fusion by apolipoprotein A-I and its amphipathic peptide analogs. J Cell Biochem

1991;45(2):224-37.

50 Wheeler RJ. The trypanolytic factor-mechanism, impacts and applications. Trends

Parasitol 2010.

51 Hajduk SL, Moore DR, Vasudevacharya J, Siqueira H, Torri AF, Tytler EM, Esko

JD. Lysis of Trypanosoma brucei by a toxic subspecies of human high density

lipoprotein. J Biol Chem 1989;264(9):5210-7.

52 Vanhollebeke B, Pays E. The trypanolytic factor of human serum: many ways to

enter the parasite, a single way to kill. Mol Microbiol 2010;76(4):806-14.

53 Vanhamme L, Paturiaux-Hanocq F, Poelvoorde P, Nolan DP, Lins L, Van den

Abbeele J, Pays A, Tebabi P, Van XH, Jacquet A, Moguilevsky N, Dieu M, Kane

JP, De BP, Brasseur R, Pays E. Apolipoprotein L-I is the trypanosome lytic factor

of human serum. Nature 2003;422(6927):83-7.

54 Raper J, Fung R, Ghiso J, Nussenzweig V, Tomlinson S. Characterization of a

novel trypanosome lytic factor from human serum. Infect Immun 1999;67(4):1910-

55 Tomlinson S, Jansen AM, Koudinov A, Ghiso JA, Choi-Miura NH, Rifkin MR,

Ohtaki S, Nussenzweig V. High-density-lipoprotein-independent killing of

Trypanosoma brucei by human serum. Mol Biochem Parasitol 1995;70(1-2):131-8.

56 Vanhollebeke B, De MG, Nielsen MJ, Pays A, Tebabi P, Dieu M, Raes M,

Moestrup SK, Pays E. A haptoglobin-hemoglobin receptor conveys innate

immunity to Trypanosoma brucei in humans. Science 2008;320(5876):677-81.

57 Samanovic M, Molina-Portela MP, Chessler AD, Burleigh BA, Raper J.

Trypanosome lytic factor, an antimicrobial high-density lipoprotein, ameliorates

Leishmania infection. PLoS Pathog 2009;5(1):e1000276.

58 Courtney HS, Zhang YM, Frank MW, Rock CO. Serum opacity factor, a

streptococcal virulence factor that binds to apolipoproteins A-I and A-II and

disrupts high density lipoprotein structure. J Biol Chem 2006;281(9):5515-21.

59 Genovese G, Friedman DJ, Ross MD, Lecordier L, Uzureau P, Freedman BI,

Bowden DW, Langefeld CD, Oleksyk TK, Uscinski Knob AL, Bernhardy AJ, Hicks

PJ, Nelson GW, Vanhollebeke B, Winkler CA, Kopp JB, Pays E, Pollak MR.

Association of trypanolytic ApoL1 variants with kidney disease in African

Americans. Science 2010;329(5993):841-5.

60 Van LS, Spillmann F, Schultheiss HP, Tschope C. High-density lipoprotein at the

interface of type 2 diabetes mellitus and cardiovascular disorders. Curr Pharm Des

2010;16(13):1504-16.

61 Mooradian AD, Haas MJ, Wong NC. Transcriptional control of apolipoprotein A-I

gene expression in diabetes. Diabetes 2004;53(3):513-20.

62 Han R, Lai R, Ding Q, Wang Z, Luo X, Zhang Y, Cui G, He J, Liu W, Chen Y.

Apolipoprotein A-I stimulates AMP-activated protein kinase and improves glucose

metabolism. Diabetologia 2007;50(9):1960-8.

63 Drew BG, Duffy SJ, Formosa MF, Natoli AK, Henstridge DC, Penfold SA, Thomas

WG, Mukhamedova N, de Court, Forbes JM, Yap FY, Kaye DM, van HG, Febbraio

MA, Kemp BE, Sviridov D, Steinberg GR, Kingwell BA. High-density lipoprotein

modulates glucose metabolism in patients with type 2 diabetes mellitus. Circulation

2009;119(15):2103-11.

64 Fryirs MA, Barter PJ, Appavoo M, Tuch BE, Tabet F, Heather AK, Rye KA. Effects

of high-density lipoproteins on pancreatic beta-cell insulin secretion. Arterioscler

Thromb Vasc Biol 2010;30(8):1642-8.

65 Brunham LR, Kruit JK, Pape TD, Timmins JM, Reuwer AQ, Vasanji Z, Marsh BJ,

Rodrigues B, Johnson JD, Parks JS, Verchere CB, Hayden MR. Beta-cell ABCA1

influences insulin secretion, glucose homeostasis and response to

thiazolidinedione treatment. Nat Med 2007;13(3):340-7.

66 Peterson SJ, Drummond G, Kim DH, Li M, Kruger AL, Ikehara S, Abraham NG. L-

4F treatment reduces adiposity, increases adiponectin levels, and improves insulin

sensitivity in obese mice. J Lipid Res 2008;49(8):1658-69.

67 Peterson SJ, Kim DH, Li M, Positano V, Vanella L, Rodella LF, Piccolomini F, Puri

N, Gastaldelli A, Kusmic C, L'Abbate A, Abraham NG. The L-4F mimetic peptide

prevents insulin resistance through increased levels of HO-1, pAMPK, and pAKT in

obese mice. J Lipid Res 2009;50(7):1293-304.

68 Ruan X, Li Z, Zhang Y, Yang L, Pan Y, Wang Z, Feng GS, Chen Y. Apolipoprotein

A-I possesses an anti-obesity effect associated with increase of energy

expenditure and upregulation of UCP1 in brown fat. J Cell Mol Med 2010.

69 de Souza JA, Vindis C, Negre-Salvayre A, Rye KA, Couturier M, Therond P,

Chantepie S, Salvayre R, Chapman MJ, Kontush A. Small, dense HDL 3 particles

attenuate apoptosis in endothelial cells: pivotal role of apolipoprotein A-I. J Cell Mol

Med 2010;14(3):608-20.

70 Fuhrman B, Gantman A, Aviram M. Paraoxonase 1 (PON1) deficiency in mice is

associated with reduced expression of macrophage SR-BI and consequently the

loss of HDL cytoprotection against apoptosis. Atherosclerosis 2010;211(1):61-8.

71 Ortiz-Munoz G, Houard X, Martin-Ventura JL, Ishida BY, Loyau S, Rossignol P,

Moreno JA, Kane JP, Chalkley RJ, Burlingame AL, Michel JB, Meilhac O. HDL

antielastase activity prevents smooth muscle cell anoikis, a potential new

antiatherogenic property. FASEB J 2009;23(9):3129-39.

72 Frias MA, Lang U, Gerber-Wicht C, James RW. Native and reconstituted HDL

protect cardiomyocytes from doxorubicin-induced apoptosis. Cardiovasc Res

2010;85(1):118-26.

73 Theilmeier G, Schmidt C, Herrmann J, Keul P, Schafers M, Herrgott I, Mersmann

J, Larmann J, Hermann S, Stypmann J, Schober O, Hildebrand R, Schulz R,

Heusch G, Haude M, von Wnuck LK, Herzog C, Schmitz M, Erbel R, Chun J,

Levkau B. High-density lipoproteins and their constituent, sphingosine-1-

phosphate, directly protect the heart against ischemia/reperfusion injury in vivo via

the S1P3 lysophospholipid receptor. Circulation 2006;114(13):1403-9.

74 Kontush A, Therond P, Zerrad A, Couturier M, Negre-Salvayre A, de Souza JA,

Chantepie S, Chapman MJ. Preferential sphingosine-1-phosphate enrichment and

sphingomyelin depletion are key features of small dense HDL3 particles: relevance

to antiapoptotic and antioxidative activities. Arterioscler Thromb Vasc Biol

2007;27(8):1843-9.

75 Rutti S, Ehses JA, Sibler RA, Prazak R, Rohrer L, Georgopoulos S, Meier DT,

Niclauss N, Berney T, Donath MY, von EA. Low- and high-density lipoproteins

modulate function, apoptosis, and proliferation of primary human and murine

pancreatic beta-cells. Endocrinology 2009;150(10):4521-30.

76 Terasaka N, Wang N, Yvan-Charvet L, Tall AR. High-density lipoprotein protects

macrophages from oxidized low-density lipoprotein-induced apoptosis by

promoting efflux of 7-ketocholesterol via ABCG1. Proc Natl Acad Sci U S A

2007;104(38):15093-8.

77 Yvan-Charvet L, Pagler TA, Seimon TA, Thorp E, Welch CL, Witztum JL, Tabas I,

Tall AR. ABCA1 and ABCG1 protect against oxidative stress-induced macrophage

apoptosis during efferocytosis. Circ Res 2010;106(12):1861-9.

78 Radojkovic C, Genoux A, Pons V, Combes G, de JH, Champagne E, Rolland C,

Perret B, Collet X, Terce F, Martinez LO. Stimulation of cell surface F1-ATPase

activity by apolipoprotein A-I inhibits endothelial cell apoptosis and promotes

proliferation. Arterioscler Thromb Vasc Biol 2009;29(7):1125-30.

79 Brodeur MR, Brissette L, Falstrault L, Moreau R. HDL3 reduces the association

and modulates the metabolism of oxidized LDL by osteoblastic cells: a protection

against cell death. J Cell Biochem 2008;105(6):1374-85.

80 de Souza JA, Vindis C, Hansel B, Negre-Salvayre A, Therond P, Serrano CV, Jr.,

Chantepie S, Salvayre R, Bruckert E, Chapman MJ, Kontush A. Metabolic

syndrome features small, apolipoprotein A-I-poor, triglyceride-rich HDL3 particles

with defective anti-apoptotic activity. Atherosclerosis 2008;197(1):84-94.

81 Mythreye K, Satterwhite LL, Davidson WS, Goldschmidt-Clermont PJ. ApoA-I

induced CD31 in bone marrow-derived vascular progenitor cells increases

adhesion: implications for vascular repair. Biochim Biophys Acta 2008;1781(11-

12):703-9.

82 Yvan-Charvet L, Pagler T, Gautier EL, Avagyan S, Siry RL, Han S, Welch CL,

Wang N, Randolph GJ, Snoeck HW, Tall AR. ATP-binding cassette transporters

and HDL suppress hematopoietic stem cell proliferation. Science

2010;328(5986):1689-93.

83 Fujimoto VY, Kane JP, Ishida BY, Bloom MS, Browne RW. High-density lipoprotein

metabolism and the human embryo. Hum Reprod Update 2010;16(1):20-38.

84 Browne RW, Bloom MS, Shelly WB, Ocque AJ, Huddleston HG, Fujimoto VY.

Follicular fluid high density lipoprotein-associated micronutrient levels are

associated with embryo fragmentation during IVF. J Assist Reprod Genet

2009;26(11-12):557-60.

85 Yesilaltay A, Morales MG, Amigo L, Zanlungo S, Rigotti A, Karackattu SL,

Donahee MH, Kozarsky KF, Krieger M. Effects of hepatic expression of the high-

density lipoprotein receptor SR-BI on lipoprotein metabolism and female fertility.

Endocrinology 2006;147(4):1577-88.

86 Nofer JR, Brodde MF, Kehrel BE. High-density lipoproteins, platelets and the

pathogenesis of atherosclerosis. Clin Exp Pharmacol Physiol 2010;37(7):726-35.

87 Dole VS, Matuskova J, Vasile E, Yesilaltay A, Bergmeier W, Bernimoulin M,

Wagner DD, Krieger M. Thrombocytopenia and platelet abnormalities in high-

density lipoprotein receptor-deficient mice. Arterioscler Thromb Vasc Biol

2008;28(6):1111-6.

88 Calkin AC, Drew BG, Ono A, Duffy SJ, Gordon MV, Schoenwaelder SM, Sviridov

D, Cooper ME, Kingwell BA, Jackson SP. Reconstituted high-density lipoprotein

attenuates platelet function in individuals with type 2 diabetes mellitus by

promoting cholesterol efflux. Circulation 2009;120(21):2095-104.

89 Kontush A, Chantepie S, Chapman MJ. Small, dense HDL particles exert potent

protection of atherogenic LDL against oxidative stress. Arterioscler Thromb Vasc

Biol 2003;23(10):1881-8.

90 Cheung MC, Albers JJ. Characterization of lipoprotein particles isolated by

immunoaffinity chromatography. Particles containing A-I and A-II and particles

containing A-I but no A-II. J Biol Chem 1984;259(19):12201-9.

91 Asztalos BF, Sloop CH, Wong L, Roheim PS. Two-dimensional electrophoresis of

plasma lipoproteins: recognition of new apo A-I-containing subpopulations.

Biochim Biophys Acta 1993;1169(3):291-300.

92 Warnick GR, Nauck M, Rifai N. Evolution of methods for measurement of HDL-

cholesterol: from ultracentrifugation to homogeneous assays. Clin Chem

2001;47(9):1579-96.

93 Duong PT, Weibel GL, Lund-Katz S, Rothblat GH, Phillips MC. Characterization

and properties of pre beta-HDL particles formed by ABCA1-mediated cellular lipid

efflux to apoA-I. J Lipid Res 2008;49(5):1006-14.

94 Haginaka JYYKM. Anion-exchange high-performance liquid chromatography

assays of plasma lipoproteins and modified low-density lipoproteins using a

ProtEx-DEAE column. J Chromatogr B Biomed Sci Appl 2001;751(1):161-7.

95 van't HF, Havel RJ. Metabolism of apolipoprotein E in plasma high density

lipoproteins from normal and cholesterol-fed rats. J Biol Chem

1982;257(18):10996-1001.

96 Stahlman M, Davidsson P, Kanmert I, Rosengren B, Boren J, Fagerberg B,

Camejo G. Proteomics and lipids of lipoproteins isolated at low salt concentrations

in D2O/sucrose or in KBr. J Lipid Res 2008;49(2):481-90.

97 Karlsson H, Leanderson P, Tagesson C, Lindahl M. Lipoproteomics I: mapping of

proteins in low-density lipoprotein using two-dimensional gel electrophoresis and

mass spectrometry. Proteomics 2005;5(2):551-65.

98 Santos RD, Schaefer EJ, Asztalos BF, Polisecki E, Wang J, Hegele RA, Martinez

LR, Miname MH, Rochitte CE, Da Luz PL, Maranhao RC. Characterization of high

density lipoprotein particles in familial apolipoprotein A-I deficiency. J Lipid Res

2008;49(2):349-57.

99 Movva R, Rader DJ. Laboratory assessment of HDL heterogeneity and function.

Clin Chem 2008;54(5):788-800.

100 Canas B, Lopez-Ferrer D, Ramos-Fernandez A, Camafeita E, Calvo E. Mass

spectrometry technologies for proteomics. Brief Funct Genomic Proteomic

2006;4(4):295-320.

101 Karlsson H, Leanderson P, Tagesson C, Lindahl M. Lipoproteomics II: mapping of

proteins in high-density lipoprotein using two-dimensional gel electrophoresis and

mass spectrometry. HDL composition 2005;5(5):1431-45.

102 Subramaniyam D, Glader P, von WK, Burneckiene J, Stevens T, Janciauskiene S.

C-36 peptide, a degradation product of alpha1-antitrypsin, modulates human

monocyte activation through LPS signaling pathways. Int J Biochem Cell Biol

2006;38(4):563-75.

103 Heller M, Stalder D, Schlappritzi E, Hayn G, Matter U, Haeberli A. Mass

spectrometry-based analytical tools for the molecular protein characterization of

human plasma lipoproteins. Proteomics 2005;5(10):2619-30.

104 Rezaee F, Casetta B, Levels JH, Speijer D, Meijers JC. Proteomic analysis of high-

density lipoprotein. Proteomics 2006;6(2):721-30.

105 Hortin GL, Shen RF, Martin BM, Remaley AT. Diverse range of small peptides

associated with high-density lipoprotein. Biochem Biophys Res Commun

2006;340(3):909-15.

106 Vaisar T, Pennathur S, Green PS, Gharib SA, Hoofnagle AN, Cheung MC, Byun J,

Vuletic S, Kassim S, Singh P, Chea H, Knopp RH, Brunzell J, Geary R, Chait A,

Zhao XQ, Elkon K, Marcovina S, Ridker P, Oram JF, Heinecke JW. Shotgun

proteomics implicates protease inhibition and complement activation in the

antiinflammatory properties of HDL. J Clin Invest 2007;117(3):746-56.

107 Green PS, Vaisar T, Pennathur S, Kulstad JJ, Moore AB, Marcovina S, Brunzell J,

Knopp RH, Zhao XQ, Heinecke JW. Combined statin and niacin therapy remodels

the high-density lipoprotein proteome. Circulation 2008;118(12):1259-67.

108 Alwaili K, Bailey D, Awan Z, Bailey SD, Ruel I, Hafiane A, Krimbou L, Laboissiere

S, Genest J. The HDL proteome in acute coronary syndromes shifts to an

inflammatory profile. Biochim Biophys Acta 2012;1821(3):405-15.

109 Watanabe J, Charles-Schoeman C, Miao Y, Elashoff D, Lee YY, Katselis G, Lee

TD, Reddy ST. Proteomic profiling following immunoaffinity capture of high-density

lipoprotein: association of acute-phase proteins and complement factors with

proinflammatory high-density lipoprotein in rheumatoid arthritis. Arthritis Rheum

2012;64(6):1828-37.

110 Boyle KE, Phillips MC, Lund-Katz S. Kinetics and mechanism of exchange of

apolipoprotein C-III molecules from very low density lipoprotein particles. Biochim

Biophys Acta 1999;1430(2):302-12.

111 Qiu X, Mistry A, Ammirati MJ, Chrunyk BA, Clark RW, Cong Y, Culp JS, Danley

DE, Freeman TB, Geoghegan KF, Griffor MC, Hawrylik SJ, Hayward CM, Hensley

P, Hoth LR, Karam GA, Lira ME, Llyod DB, Mcgrawth KM, Stutzman-Engwall KJ,

Subashi AK, Subashi TA, Thompson JF, Wang I, Zhao H, Seddon AP. Crystal

structure of cholesteryl ester transfer protein reveals a long tunnel and four bound

lipid molecules. Nat Struct Mol Biol 2007;14:106-13.

112 Asztalos BF, Schaefer EJ. High-density lipoprotein subpopulations in pathologic

conditions. Am J Cardiol 2003;91(7A):12E-7E.

113 Davidson WS, Silva RA, Chantepie S, Lagor WR, Chapman MJ, Kontush A.

Proteomic analysis of defined HDL subpopulations reveals particle-specific protein

clusters: relevance to antioxidative function. Arterioscler Thromb Vasc Biol

2009;29(6):870-6.

114 Jonas A. Lecithin-cholesterol acyltransferase in the metabolism of high- density

lipoproteins. [Review]. Biochim Biophys Acta 1991;1084(3):205-20.

115 Wang X, Driscoll DM, Morton RE. Molecular cloning and expression of lipid

transfer inhibitor protein reveals its identity with apolipoprotein F. J Biol Chem

1999;274(3):1814-20.

116 Jahangiri A, Rader DJ, Marchadier D, Curtiss LK, Bonnet DJ, Rye KA. Evidence

that endothelial lipase remodels high density lipoproteins without mediating the

dissociation of apolipoprotein A-I. J Lipid Res 2005;46(5):896-903.

117 Rifkin MR. Identification of the trypanocidal factor in normal human serum: high

density lipoprotein. Proc Natl Acad Sci U S A 1978;75(7):3450-4.

118 Nielsen MJ, Nielsen LB, Moestrup SK. High-density lipoprotein and innate

immunity. Future Lipidology 2006;1(6):729-34.

119 Shiflett AM, Bishop JR, Pahwa A, Hajduk SL. Human high density lipoproteins are

platforms for the assembly of multi-component innate immune complexes. J Biol

Chem 2005;280(38):32578-85.

120 Gordon SM, Deng J, Lu LJ, Davidson WS. Proteomic characterization of human

plasma high density lipoprotein fractionated by gel filtration chromatography. J

Proteome Res 2010;9:5239-49.

121 Maron DJ, Fazio S, Linton MF. Current perspectives on statins. Circulation

2000;101(2):207-13.

122 Ansell BJ, Navab M, Hama S, Kamranpour N, Fonarow G, Hough G, Rahmani S,

Mottahedeh R, Dave R, Reddy ST, Fogelman AM. Inflammatory/antiinflammatory

properties of high-density lipoprotein distinguish patients from control subjects

better than high-density lipoprotein cholesterol levels and are favorably affected by

simvastatin treatment. Circulation 2003;108(22):2751-6.

Figure 1-1. The increasing functional heterogeneity of high density lipoprotein. Numerous recent studies have begun to uncover HDL functions that vary surprisingly from traditionally recognized lipid transport roles. Further exploration of these functions will be useful for understanding the potential of HDL as a treatment option for non-cardiovascular pathologies.

Table 1-1. Master compendium of HDL associated proteins identified using modern proteomic strategies (arranged alphabetically by recommended protein name, proteins in red appeared in at least 3 different MS studies or had independent biochemical evidence of HDL residence)

Recommended protein name Abbrev. Acc. # Vaisar (1) Rezaee (2) Karlsson (3) Heller (4) Hortin (5) Davidson (6) Gordon (7) Alwaili (8) Watanabe (9) Hits* Likely# afamin AFM P43652 1 1 2 0 albumin ALB P02768 1 1 1 1 1 5 1 alpha-1-acid glycoprotein 1 ORM1 P02763 1 1 2 0 alpha-1-acid glycoprotein 2 ORM2 P19652 1 1 1 1 4 1 alpha-1-antichymotrypsin SERPINA3 P01011 1 1 2 0 alpha-1-antitrypsin SERPINA1 P01009 1 1 1 1 1 1 1 7 1 alpha-1B-glycoprotein A1BG P04217 1 1 1 1 1 5 1 alpha-2 antiplasmin SERPINF2 P08697 1 1 1 3 1 alpha-2-HS-glycoprotein AHSG P02765 1 1 1 1 1 5 1 aminopeptidase N ANPEP P15144 1 1 0 angiotensinogen AGT P01019 1 1 1 3 1 antithrombin-III SERPINC1 P01008 1 1 1 3 1 apolipoprotein A-I APOA1 P02647 1 1 1 1 1 1 1 1 1 9 1 apolipoprotein A-II APOA2 P02652 1 1 1 1 1 1 1 1 8 1 apolipoprotein A-IV APOA4 P06727 1 1 1 1 1 1 1 1 1 9 1 apolipoprotein A-V APOA5 Q6Q788 0 1 apolipoprotein B APOB P04114 1 1 1 1 1 5 1 apolipoprotein C-I APOC1 P02654 1 1 1 1 1 1 1 7 1 apolipoprotein C-II APOC2 P02655 1 1 1 1 1 1 1 7 1 apolipoprotein C-III APOC3 P02656 1 1 1 1 1 1 1 1 1 9 1 apolipoprotein C-IV APOC4 P55056 1 1 1 3 1 apolipoprotein D APOD P05090 1 1 1 1 1 1 6 1 apolipoprotein E APOE P02649 1 1 1 1 1 1 1 1 8 1 apolipoprotein F APOF Q13790 1 1 1 1 1 5 1 apolipoprotein H APOH P02749 1 1 1 1 4 1 apolipoprotein J CLU P10909 1 1 1 1 1 1 1 7 1 apolipoprotein M APOM O95445 1 1 1 1 1 1 1 7 1 apolipoprotein O APOO Q9BUR5 0 1 apolipoprotein(a) LPA P08519 1 1 2 0 apolipoprotien L1 APOL1 O14791 1 1 1 1 1 1 1 1 1 9 1 beta-2-microglobulin B2M P61769 1 1 0 bifunctional proten NCOAT MGEA5 O60502 1 1 0 C4b-binding protein alpha chain C4BPA P04003 0 0 carboxypeptidase N CPN P15169 1 1 0 cathelicidin antimicrobial peptide CAMP P49913 1 1 0 ceruloplasmin CP P00450 1 1 0 cholesteryl ester transfer protein CETP P11597 1 1 2 1 Cip1-interacting zinc finger protein CIZ1 Q9ULV3 1 1 0 coagulation factor V F5 P12259 1 1 0 coagulation factor XII F12 P00748 1 1 0 complement C1q subcomponent subunit B C1QB P02746 1 1 0 complement C1q subcomponent subunit C C1QC P02747 1 1 0 complement C1r C1R P00736 1 1 0 complement C1s subcomponent C1S P09871 1 1 1 3 1 complement C2 C2 P06681 1 1 2 0 complement C3 C3 P01024 1 1 1 1 4 1 complement C4-A C4A P0C0L4 1 1 0 complement C4-B C4B P0C0L5 1 1 1 1 4 1 complement C5 C5 P01031 1 1 0 complement C8-beta C8B P07358 1 1 0 complement C9 C9 P02748 1 1 2 0 complement factor B CFB P00751 1 1 1 3 1 complement factor H CFH P08603 1 1 2 0 C-reactive protein CRP P02741 1 1 0 C-type lectin domain family 3 member A CLEC3A O75596 1 1 0 desmocollin-1 DSC1 Q08554 1 1 0 fibrinogen alpha chain FGA P02671 1 1 1 1 1 1 1 7 1 fibrinogen beta chain FGB P02675 1 1 2 0 fibrinogen gamma chain FGG P02679 1 1 0 fibronectin FN1 P02751 1 1 2 0 ficolin-3 FCN3 O75636 1 1 0 galectin-7 LGALS7 P47929 1 1 0 gelsolin GSN P06396 1 1 2 0 growth arrest-specific protein 6 GAS6 Q14393 1 1 0 haptoglobin HP P00738 1 1 0 haptoglobin related protein HPR P00739 1 1 1 1 1 1 1 7 1 hemoglobin subunit alpha HBA1 P69905 1 1 2 0 hemoglobin subunit beta HBB P68871 1 1 2 0 hemopexin HPX P02790 1 1 1 1 4 1 heparin cofactor 2 SERPIND1 P05546 1 1 1 3 1 histidine-rich glycoprotein HRG P04196 1 1 0 histone H2A HIST1H2AG P0C0S8 1 1 0 HLA-A protein (fragment) HLA-A Q29946 1 1 0 hyaluronan BP 2 HABP2 Q14520 1 1 0 insulin-like growth factor-binding protein 1 IGFBP1 P08833 1 1 0 insulinoma-associated protein 1 INSM1 Q01101 1 1 0 inter alpha trypsin inhibitor 1 ITIH1 P19827 1 1 2 0 inter alpha trypsin inhibitor 2 ITIH2 P19823 1 1 1 1 1 5 1 inter alpha trypsin inhibitor 3 ITIH3 Q06033 1 1 0 inter alpha trypsin inhibitor 4 ITIH4 Q14624 1 1 1 1 1 5 1 isocitrate dehydrogenase [NAD] subunit alpha IDH3A P50213 1 1 0 kallistatin SERPINA4 P29622 1 1 0 kininogen-1 KNG1 P01042 1 1 1 3 1 latent-transforming growth factor beta-binding protein 1LTBP1 Q14766 1 1 0 latent-transforming growth factor beta-binding protein 2LTBP2 Q14767 1 1 0 lecithin:cholesterol acyltransferase LCAT P04180 1 1 2 1 leucine-rich alpha-2-glycoprotein LRG1 P02750 1 1 0 lipopolysaccaride binding protein LBP P18428 1 1 2 1 lipoprotein lipase LPL P06858 1 1 0 lumican LUM P51884 1 1 2 0 macrophage colony-stimulating factor 1 CSF1 P09603 1 1 0 n-acetylmuramoyl-L-alanine amidase PGLYRP2 Q96PD5 1 1 0 Neurogenic locus notch homolog protein 1 NOTCH1 P46531 1 1 0 paraoxonase 1 PON1 P27169 1 1 1 1 1 1 1 7 1 paraoxonase 3 PON3 Q15166 1 1 1 3 1 phosholipid transfer protein PLTP P55058 1 1 1 3 1 phosphatidylinositol-glycan-specific phospholipase D GPLD1 P80108 1 1 0 pigment epithelium-derived factor SERPINF1 P36955 1 1 1 3 1 plasma kallikrein KLKB1 P03952 1 1 2 0

55 plasma protease C1 inhibitor SERPING1 P05155 1 1 1 3 1 plasma serine proteinase inhibitor SERPINA5 P05154 1 1 0 plasminogen PLG P00747 1 1 0 platelet activating factor aryl hydrolase PAFAH1B1 P43034 1 1 1 platelet basic protein PPBP P02775 1 1 0 POU 5 domain protein POU5F2 Q8N7G0 1 1 0 pregnancy-specific beta-1- glycoprotein PSG1 P11464 1 1 2 0 prenylcystein oxidase PCYOX1 Q9UHG3 1 1 2 0 protein AMBP AMBP P02760 1 1 1 3 1 prothrombin F2 P00734 1 1 1 1 4 1 retinol binding protein 4 RBP4 P02753 1 1 1 3 1 ryanodine receptor 2 RYR2 Q92736 1 1 0 serotransferrin TF P02787 1 1 1 1 1 1 6 1 serum amyloid A1 &2 SAA1/SAA2 P02735 1 1 1 1 1 5 1 serum amyloid A4 SAA4 P35542 1 1 1 1 1 1 6 1 serum amyloid P APCS P02743 1 1 2 1 SH2 domain-containing protein 1A SH2D1A O60880 1 1 0 sialic acid-binding Ig-like lectin 5 SIGLEC5 O15389 1 1 0 T-cell surface glycoprotein CD5 CD5 P06127 1 1 0 tetranectin CLEC3B P05452 1 1 0 tissue factor pathway inhibitor TFPI P10646 1 1 0 transthyretin TTR P02766 1 1 1 1 1 1 1 1 8 1 vitamin D binding protein VTDB P02774 1 1 1 1 1 5 1 vitamin K dependent protein S PROS1 P07225 1 1 0 vitronectin VTN P04004 1 1 1 1 4 1 zinc-alpha-2-glycoprotein AZGP1 P25311 1 1 0 α-2-macroglobulin A2M P01023 1 1 2 0 α-amylase 1 AMY1A P04745 1 1 0 * The term "hits" refers to the number of cited mass spectrometry studies that found a particular protein. # Proteins likely to be truly HDL associated are given a score of 1 and shown in red. They are considered "likely" because the were: Total "hits": 127 1) Identified in at least 3 of the MS studies cited above from different laboratories, or 2) Identified as associtated with HDL by other biochemical means (immunoprecipitation, binding studies, etc). "likely" HDL proteins*: 55

Mass spectrometry studies cited in the table: 1) Vaisar et al., J. Clin. Invest. 2007; 117(3): 746-756. Used LC-MS and spectral counting on total HDL (20 subjects) and HDL3 (6 controls and 7 CAD subjects) isolated by ultracentrifugation. 2) Rezaee et al., Proteomics 2006; 6: 721-730. Used 1 and 2 dimensional electrophoresis with MALDI-TOF on total HDL isolated by ultracentrifugation and by immuno-isolation. 3) Karlsson et al., Proteomics 2005; 5; 1431-1445. Used 2 dimensional electrophoresis with MALDI-TOF on HDL2 and HDL3 isolated by ultracentrifugation. 4) Heller et al., Proteomics 2005; 5; 2619-2630. Used LCMS and MALDI to study human plasma fractionated by ultracentrifugation. 5) Hortin et al., Biochemical and Biophysical Research Communications. 2006; 340;909-915. Used MALDI-TOF to analyze small peptides from total human HDL isolated by ultracentrifugation. 6) Davidson et al., Arteriosclerosis Thrombosis and Vascular Biology. 2009; 29:870-876. Used shotgun LC-MS to study 5 density subfractions of human HDL. 7) Gordon et al., Journal of Proteome Research. 2010; 9; 5239-5249. Used shotgun LC-MS to study 17 fractions of human plasma using a lipid binding agent to distinguish lipoproteins. 8) Alwaili et al., Biochimica et Biophysica Acta. 2012; 1821; 405-415. Used 1 D electrophoresis and LC-MS to study total human HDL isolated by ultracentrifugation. 9) Watanabe et al., Arthritis and Rheumatism. 2012; 64; 1828-1837. Used LC-MS to study human plasma HDL obtained by immunoaffinity capture.

Chapter 2: Proteomic characterization of human plasma high density lipoprotein fractionated by gel filtration chromatography

Introduction

Lipoproteins are dynamic particles composed of lipid and proteins called apolipoproteins 1. They are formed in the liver, small intestine and certain macrophages and are secreted into the bloodstream where they mediate transport and metabolism of lipids. Chylomicrons (CM), very low and low density lipoproteins (VLDL/LDL) act in the delivery of dietary or liver-derived triglycerides and cholesterol to peripheral tissues for use or storage, while HDL are responsible for RCT 2 (described in Chapter 1). High plasma levels of LDL and VLDL have been correlated with increased risk for CVD but

HDL levels are inversely correlated 3, 4. While HDL's role in RCT undoubtedly plays a major role in cardio-protection, recent studies indicate that HDL also possesses antioxidative 5 and anti-inflammatory 6, 7 properties which likely contribute to its cardioprotective effects.

To identify a mechanistic basis for these observed functions, research quickly focused on HDL apolipoproteins 8. HDL are secreted from the liver as nascent phospholipid ‘discs’ made stable by their association with the most abundant HDL protein apolipoprotein A-I (apoA-I). As these particles accumulate free cholesterol from peripheral tissues, HDL associated lecithin cholesterol acyl transferase (LCAT) esterifies fatty acids to free cholesterol to form cholesteryl esters which accumulate as a

57 hydrophobic core in the particle, eventually resulting in a spherical morphology. Mature

HDL can transfer cholesteryl esters to LDL in exchange for triglyceride via another HDL protein, cholesteryl ester transfer protein (CETP). Other proteins with roles outside of

RCT have been identified on HDL. For example, paraoxonase-1 (Pon1) which may be responsible for HDL’s ability to prevent oxidation of LDL particles 9; oxidized LDL are a major contributing factor to atherosclerotic development 10, 11.

As described in the previous chapter, several groups have applied mass spectrometry-based proteomic technologies to identify nearly 100 protein components of HDL 12, though there is only substantial agreement among studies on about 55 of these (Table 1-1). Many of these newly identified HDL associated proteins mediate functions that are surprisingly outside the realm of lipid transport and metabolism. For example, HDL was found to be a host for numerous protease inhibitors as well as mediators of the complement cascade13, suggesting possible roles for HDL in innate immunity. This raises interesting new possibilities for functional roles of HDL and suggests many more remain to be discovered.

Despite these advances, a full proteomic understanding of HDL remains incomplete. To date, nearly all mass spectrometry-based proteomics studies of HDL have utilized density gradient ultracentrifugation (UC) based methods for the isolation of

HDL from human plasma. This method is optimal for MS-based proteomics as it has the advantage of quantitatively floating the relatively light lipid-containing proteins away from the dense non-lipoprotein associated proteins. However, the separation involves high shear forces and prolonged exposure to elevated salt concentrations that likely alter HDL functionality and composition. For example, Van’t Hooft et al. showed that

58 about half of the apoE on HDL is dissociated during UC and, compared with gel filtration isolated HDL, UC isolated HDL interacted more avidly with the apoE receptor due to either changes in apoE conformation or depletion of other apolipoproteins 14. Thus, there is a significant need to analyze the HDL proteome using alternative separation techniques. An attractive alternative is gel filtration chromatography, which separates plasma components by size. The separation can be performed quickly under physiological salt and shear conditions and thus is less likely to alter the HDL proteome.

However, the significant disadvantage of this technique (and most other non-UC techniques) is the overlap between HDL and many high abundance plasma proteins.

For example, abundant immunoglobulins can have a MW in the range of 150 - 900 kDa

(depending on class) which overlaps with the 150 – 360 kDa mass range of most HDL particles. In addition, the presence of extremely high abundance small proteins, such as human albumin (40-60 mg/mL), can significantly contaminate HDL fractions. These contaminants can reduce MS detection of the desired HDL proteins due to ion suppression effects, or by forcing the instrument to devote the majority of its duty cycle to the MS/MS analysis of peptide ions derived from abundant, contaminant proteins.

These issues have been major roadblocks to the use of non-centrifugal methods for

HDL proteomic analysis.

In this study, we approached this problem in two ways. First, we derived gel filtration conditions that separate HDL from the bulk of high abundance low MW proteins such as albumin. Then we developed an affinity technique to specifically analyze only those proteins that are associated with plasma phospholipids. As a result, we have identified several new HDL associated proteins, provided more evidence that

59 ultracentrifugal separations of HDL may modify its proteome, and demonstrated distinct distribution patterns for a variety of proteins between the different size HDL particles.

Experimental

Plasma collection. Venous blood was collected from fasted (≥12 hours), apparently healthy normolipidemic (total cholesterol between 125-200 mg/dL; HDL-C ≥ 40 mg/dL; triglycerides < 150 mg/dL) male donors (ages: 21, 22 and 34) by a trained phlebotomist using BD Vacutainer® Plus Plastic Citrate Tubes containing buffered sodium citrate

(0.105 M) as an anticoagulant. Cellular components were pelleted by centrifugation at

~1,590 x g for 15 minutes in a Horizon mini-E (Quest Diagnostics) at room temperature.

Plasma was stored at 4ºC until gel filtration separation, always within 16 hours.

Samples were never frozen.

Plasma separation by gel filtration chromatography. 370 µL of plasma from a single subject was applied directly to three Superdex 200 gel filtration columns (10/300 GL;

GE Healthcare) arranged in series on an ÄKTA™ FPLC system (GE Healthcare). The sample processed at a flow rate of 0.3 mL/min in standard Tris buffer (STB) (10mM Tris,

0.15M NaCl, 1mM EDTA, 0.2% NaN3). Eluate was collected as forty-seven 1.5 mL fractions on a Frac 900 fraction collector (GE healthcare) maintained at 4ºC. Each fraction was assessed for protein, phospholipid and total cholesterol by colorimetric kits from Wako (Richmond, VA). For ether delipidation protein shift experiments, plasma (5 mL) was delipidated with butanol-di-isopropyl ether (40:60, 10 mL) according to a

60 procedure described by Cham and Knowles 15. The volume of freshly delipidated plasma was then adjusted with STB to match the protein concentration of normal plasma and applied to triple Superdex 200 columns exactly as with normal plasma.

Fractions collected from delipidated plasma were not subjected to LRA treatment

(described below).

Purification of phospholipid-containing particles using calcium silicate hydrate

(LRA). To isolate lipoprotein particles from co-eluting proteins in the collected fractions, we used a commercially available synthetic calcium silicate hydrate called Lipid

Removal Agent (Supelco). This compound, developed for the removal of lipids in biopharmaceutical production, tightly binds lipids and lipoproteins. In a centrifuge tube,

45 µg of LRA (from 100 mg/mL stock solution in 50 mM ammonium bicarbonate) per

1µg of PL in 400 µL of fraction were mixed gently for 30 minutes at room temperature.

The LRA was then pelleted by centrifugation (~2200 x g for 2 min.) in a minicentrifuge

(Fisher) and the supernatant containing lipid-free plasma proteins was removed. The

LRA was then washed with 50 mM ammonium bicarbonate (AB). All PL-containing fractions from each subject’s FPLC separation were carried through this process individually.

Western blotting for apoA-I. Purified human apoA-I, UC isolated HDL, supernatant from LRA procedure, and SDS elution were run on 4-15% PAGE, then transferred to a

PVDF membrane. Membranes were probed with rabbit anti-human apoA-I antibody

(Calbiochem, 178422).

Mass spectrometry analysis of fractions. HDL particles were subjected to trypsin digestion while still bound to the LRA. 1.5 µg of sequencing grade trypsin (Promega) in

25 µL of 50 mM AB was added to each LRA pellet and incubated at 37ºC overnight on a rotating plate. To collect the digested peptides, the LRA was washed with 125 µL of 50 mM AB. Peptides were first reduced and then carbamidomethylated with dithiothreitol

(200mM; 30 min. at 37ºC) and iodoacetamide (800mM; 30 min. at room temperature), respectively. Peptide solutions were then lyophilized to dryness and stored at -20ºC until analyzed by mass spectrometry. Dried peptides were reconstituted in 15 µL of

0.1% formic acid in water. An Agilent 1100 series Autosampler/HPLC was used to draw

0.5 µL of sample and inject it onto a C18 reverse phase column (GRACE; 150 x

0.500mm) where an acetonitrile concentration gradient (5-30% in water with 0.1% formic acid) was used to elute peptides for on-line ESI-MS/MS by a QStar XL mass spectrometer (Applied Biosystems). Column cleaning was performed automatically with

2 cycles of a 5-85% acetonitrile gradient lasting 15 min. each between runs.

MS data analysis. To identify the protein composition of particles contained in the various gel filtration fractions, peak lists generated from analysis of each fraction were scanned against the UniProtKB/Swiss-Prot Protein Knowledgebase (release 57.0,

03/2009) using both the Mascot (version 2.1) and X!Tandem (version 2007.01.01.1) search engines. Search criteria included: human taxonomy, variable modifications of

Met oxidation and carbamidomethylation, both peptide tolerance and MS/MS tolerance were set to ± 0.15 Da, and up to 3 missed tryptic cleavage sites were allowed. Scaffold

62 software (version Scaffold_2_04_00, Proteome Software) was used to validate MS/MS based peptide and protein identifications. Peptide identification required a value of 90% probability (using data from both Mascot and X!Tandem) using the Peptide Prophet algorithm 16. Positive protein identification also required a value of 90% probability by the Protein Prophet algorithm 17. Also, a minimum of 2 peptides were required unless the protein in question was found with single peptide hits in multiple consecutive fractions that were consistent across all subjects. Since equal volumes of sample were applied to the MS analysis, not equal protein contents, we reasoned that the relative amount of a given protein present in a given fraction should be proportional to the number of spectral counts (i.e. the number of MS/MS spectra assigned to a particular protein) in each fraction. In no case were conclusions made about the relative abundance of two different proteins on the basis of peptide counting. We have previously demonstrated that this approach provides a semi-quantitative abundance pattern across the fractions that matches well with patterns derived from immunological analyses 18.

Results

Optimization of gel filtration resolution for human plasma HDL. A widely used method for gel filtration-based separations of plasma lipoproteins involves the application of plasma to two Superose 6 columns (GE Healthcare) connected in series.

This method has proven useful in analysis of the lipid content of plasma lipoproteins 19.

We began by evaluating this method in terms of its effectiveness in separating HDL

63 proteins from the high abundance, non-lipidated proteins such as albumin. Figure 2-1a shows the protein and total cholesterol distribution of human plasma separated by the tandem Superose 6 protocol. The cholesterol peaks 1, 2 and 3 represent VLDL, LDL and HDL respectively. While the lipid peaks were well distinguished, it is clear that the

HDL cholesterol peak underwent substantial overlap with the majority of plasma proteins (peak 4). To optimize separation of HDL from the abundant protein peak, we developed a method that utilized three Superdex 200 columns arranged in series (Fig.

2-1b). In addition to the extra resolving power contributed by a third column, the

Superdex 200 matrix pore size distribution allowed for greater resolution within the HDL size range. In this case, VLDL and LDL cholesterol ran together (peak 1 & 2) while HDL cholesterol distributed in a broad peak 3. It is clear that this protocol separates the majority of HDL cholesterol from the free plasma proteins evident in peak 4. Figure 2-2 shows an SDS-PAGE analysis that compared total human HDL isolated by ultracentrifugation vs. a pooled sample of gel filtered HDL, volumes 30 – 35 mL in Fig.

2-1b. The major HDL proteins apoA-I and apoA-II are visible in both samples.

However, the GF sample still contained human albumin. We also observed higher molecular weight proteins in the GF sample that may or may not be truly associated with the HDL particles. Since further optimizations of the gel filtration protocol failed to significantly improve HDL separation from albumin, we elected to pursue methods that would allow affinity isolation of those proteins that were specifically associated with lipid.

Selection of lipid bound proteins using calcium silicate hydrate. After exploring a number of potential strategies for isolating lipid-bound proteins, we developed a method

64 that utilizes a commercially available hydrated calcium silicate resin (LRA) with a high binding affinity for phospholipid 20. In optimization experiments, we determined that 150

µg of LRA could bind about 1 µg of plasma phospholipid in STB at pH 8.0. We first tested the ability of LRA to sequester HDL and LDL that had been previously purified by ultracentrifugation. For both HDL and LDL, exposure to LRA resulted in the loss of

99.9% of both phospholipid and cholesterol in the flow-through, indicating the quantitative binding of both lipoproteins to the resin (Table 2-1). Additionally, the major protein components of both LDL (apoB, Fig. 2-3a) and HDL (apoA-I, Fig. 2-3b), were removed from solution after LRA (lanes 1 and 2) and were recovered from the resin with an SDS wash (lane 3). Figure 2-3c shows the HDL experiment as analyzed by Western blot using an anti-apoA-I antibody, confirming that apoA-I was almost completely removed from the supernatant. We next assessed the capacity of LRA to remove phospholipid-containing particles from fractions produced by gel filtration chromatography. Figure 2-4a compares the phospholipid profile of plasma separated by the triple Superdex protocol before and after the fractions were treated with LRA.

LRA was capable of quantitatively removing nearly all PL associated with the VLDL/LDL peak as well as the major HDL peak. Interestingly, it failed to bind a small amount of PL that co-migrates with plasma proteins. This may represent sequestered phospholipids or lyso-PC's that are tightly associated with small proteins. SDS PAGE revealed that the LRA specifically removed apoA-I from the supernatant while allowing albumin and several other proteins to wash through (Fig. 2-4b, lanes 1 and 2). When the resin was boiled in sample buffer and analyzed by SDS PAGE, it is clear that apoA-I and not albumin was retained on the resin (Fig. 2-4b, lane 3). Similarly, the same analysis

65 performed with fraction 16 in the VLDL/LDL range showed that apoB (Fig. 2-4c, band at the top of the gel) was also selectively retained on the resin.

Given that PL binds extremely tightly to the resin by an unknown mechanism, we have not identified a practical way to recover LRA-bound lipoproteins intact, at least not in a manner consistent with subsequent mass spectrometry analysis. However, we found that we could trypsinize the particles while still in contact with the resin. The resulting peptides were then eluted from the resin and used for MS analysis. Since the lipid remains associated with the resin, it was not necessary to perform subsequent delipidation steps prior to the MS analysis. The obvious disadvantage of this approach is the possibility that certain hydrophobic peptides may remain associated with the lipid after trypsinization. However, we found that overall peptide detection and sequence coverage of most of the lower abundance HDL proteins analyzed by the LRA method was comparable to those isolated by UC without the LRA step (Appendix 2: Supplement table 1).

ESI-MS/MS analysis of lipid associated proteins. Phospholipid associated proteins in each PL-containing fraction were identified by LC-ESI-MS/MS and subsequent database searching using criteria described in the experimental section. Our analysis identified 81, 98 and 103 proteins across all fractions for the three subjects studied, of these, 79 were common across all subjects. Many of these proteins had been shown to associate with UC-isolated HDL in previous studies, however many were potentially new phospholipid-associated proteins.

To evaluate the potential for phospholipid independent (i.e. non-specific) binding to the resin, we performed a set of experiments where human plasma was first subjected to an ether based delipidation procedure shown to cause minimal protein denaturation15, prior to separation on the columns. The rationale was that lipid- associated proteins will migrate with a different apparent size after the lipid is removed.

However, those that are not lipid-associated, but bind to LRA non-specifically, should elute at the same volume. By monitoring shifts in protein elution patterns in delipidated vs. control samples, we distinguished proteins that were most likely associated with lipid. An example of each case is shown in Figure 2-5. Plasminogen is a common plasma zymogen, the precursor for plasmin, an enzyme responsible for dissolving blood clots. This protein is not known to be associated with lipid and its elution patterns failed to shift when plasma was delipidated (Fig. 2-5a). However, a protein with an established association to HDL particles, complement component C3, showed a dramatic change in its elution pattern when plasma was delipidated prior to separation

(Fig. 2-5b). All previously known HDL associated proteins identified in this study were found to undergo some degree of elution profile shift in response to delipidation.

Moreover, apolipoprotein B, the primary protein component of LDL, showed similar behavior. Of the 79 proteins that passed our identification criteria for all three subjects studied, 43 proteins exhibited significant elution volume shifts in two independent delipidation experiments. Four proteins whose association with HDL has been previously established were detected but at levels too low to determine if a definitive shift had occurred. These were included in the analysis giving a total of 47 lipid associated proteins. Those proteins that failed to undergo a profile shift (i.e. those that

67 bind the resin via mec hanisms independent from PL) are shown in Appendix 2:

Supplement Table 2. The lipid-associated proteins are shown in Figure 2-6 along with the sum of their peptides across all PL-containing fractions collected from individual triple Superdex runs on all 3 subjects. Of the 47 proteins identified as lipid associated,

17 are newly identified as being associated with lipidated particles in plasma; these proteins are indicated with asterisks in Figure 2-6. Of these 17 proteins, 14 were found to elute within fractions 19-29, which is where the majority of the apoA-I elutes. These fractions likely correspond to “HDL” as traditionally defined by gradient ultracentrifugation (see Discussion). The other 3 proteins, complement C1q subcomponent subunits B and C, and ficolin-3, co-migrate with LDL/VLDL sized particles in fractions 13-18.

Using HDL subfractions defined by density ultracentrifugation, we previously showed that HDL associated proteins can be grouped into different classifications depending on their distribution patterns between dense and light fractions of HDL 18.

Figure 2-7 displays the relative distribution of selected common HDL proteins across the gel filtration fractions while the elution patterns of all detected lipid-associated proteins are shown as a heat map in Fig. 2-8. Larger apoB containing lipoproteins eluted in earlier fractions. ApoB protein abundance peaked at fraction 16 while apoA-I protein abundance peaked later in fraction 24. These protein distribution patterns correlated with the PL peaks indicated as “LDL” or “HDL” in Fig. 2-4. The major HDL proteins, apoA-I and apoA-II, were found in nearly all PL-containing fractions. Interestingly, the other lipid associated proteins distributed across fractions in distinct patterns. For example, several of the complement proteins identified seemed to be grouped primarily

68 in fractions 22-24 while apoA-IV was concentrated to the smallest particles eluting in fractions 27 and 28.

Discussion

HDL is, by definition, distinguished in terms of particle density as originally exploited for separation by gradient ultracentrifugation 21. Thus, one immediately encounters a nomenclature issue when attempting to separate these particles by methods that do not rely on density. In such case, it is tempting to define HDL on the basis of its major protein apoA-I. However, it is clear from Figs. 2-7, 2-8 and previous studies 18 that apoB containing lipoproteins such as LDL also contain significant amounts of apoA-I. In this study, we have separated fresh plasma across a broad size range that includes the traditional VLDL, LDL and HDL particle sizes. By treating all fractions with a phospholipid binding agent, we have technically measured proteins that are associated with plasma phospholipid, rather than any specific lipoprotein class.

Nevertheless, in order to relate these gel filtration results to traditional definitions of

HDL, we elected to use the presence of apoB, the core constituent of LDL as the key distinguisher. We therefore defined fractions 14-18 as the VLDL/LDL fraction due to the presence of apoB. The remaining fractions 19-29 were considered "HDL", though it is recognized that GF and UC isolate overlapping, but possibly distinct sets of particles.

We suggest these particles might be better referred to as PL-rich lipoproteins.

A recent study reported a proteomic analysis of fractions of human plasma collected by gel filtration on a single Superdex 200 column 22. This study identified the

69 majority of known HDL associated proteins, but made no attempt to distinguish HDL- associated proteins from the multitude of abundant plasma proteins which co-elute. In this work we have overcome a major barrier standing in the way of using non-UC based methods to separate human plasma HDL for proteomic analysis. The use of the calcium silicate hydrate combined with optimized gel filtration conditions resulted in the identification of some 14 new proteins that associate with phospholipid in the HDL fractions.

We examined the Biological Process and Molecular Function Gene Ontology

(GO) annotations for the 17 newly lipid associated proteins as well as the 30 previously identified HDL associated proteins found in this study. Our results were consistent with previous HDL proteome studies but also pointed out some potentially new functions

(Figure 2-9). For each annotated function, we calculated its enrichment among either previously known (# of proteins with given function / 30) or the newly identified (# of proteins with given function / 17) lipid associated proteins found in this study. The statistical significance is given by p-values in parenthesis. In addition to those identified by Vaisar et al. 13, we have identified 8 phospholipid associated proteins with known functions in the complement cascade. Three of these were found to co-migrate with apoB containing lipoproteins - complement C1q subcomponent subunits B and C which function in activation of the classical pathway 23 and ficolin-3 which is involved in complement activation via the lectin pathway 24. The remainder, complement C1s, C2,

C5, factor B and plasma protease C1 inhibitor were distributed across the HDL containing fractions. The addition of these proteins to HDL's repertoire further implicates the lipoprotein class in the complement pathway and innate immunity. This

70 study also confirmed 6 other complement proteins previously described in Vaisar's study, bringing the total count of HDL associated proteins with roles in complement function to about 14. Interestingly, plasma protease C1 inhibitor also plays a role in blood coagulation along with heparin cofactor 2 and antithrombin III. Isolated HDL have been found to possess anticoagulant properties 25, however the physical basis for this is not well understood. The presence of these proteins on HDL may begin to explain this observation.

Others have identified several HDL proteins belonging to the serine protease inhibitor (SERPIN) superfamily 13, 18, 26, 27. In this study we have identified 2 previously unreported SERPIN proteins. The first, alpha-1-antichymotrypsin (AACT) is an acute phase protein secreted by the liver under inflammatory conditions and has inhibitory activity against several proteases 28, 29. The second, pigment epithelium derived factor

(PEDF), is a member of the serine protease superfamily but has no known protease inhibitor activity 30. GO analysis annotated this protein as a positive regulator of neurogenesis, promoting the development and maintenance of motor neurons 31.

Recently, a causal role for PEDF in obesity induced insulin resistance has been identified 32.

The remaining newly identified proteins were linked to functions which have not yet been attributed to HDL. Known functions include skeletal development (tetranectin)

33, collagen fibril organization and visual perception (lumican) 34, 35, and signal transduction (insulin like growth factor binding protein acid labile subunit; ALS) 36. The latter is a liver secreted protein whose function is to bind to, and increase the half-life of, insulin like growth factor (IGF) in the plasma. Humans deficient in ALS exhibit

71 decreased levels of plasma IGF-I, IGF-II and IGF binding protein 3 due to increased clearance 37. The reason for HDL localization of these proteins is not yet apparent and invites future study.

In addition to discovering new HDL-associated proteins, the increased fractionation potential of GF has allowed us to visualize HDL protein distribution across particle size with unprecedented resolution. Figure 2-8 shows that most of the identified proteins were distributed in distinct patterns across the different sized particles. In a previous study, we separated human HDL into only 5 density subfractions by UC and also found that individual HDL proteins can be grouped with respect to their distributions among the subfractions18. In that study, we found that many of the lower abundance proteins tended to cluster in the densest subfractions which are generally suggestive of a smaller particle size. The results of the current study confirmed many of these classifications. For example, apoA-IV, alpha-1-antitrypsin and transthyretin were found exclusively in the densest HDL3c fractions by UC and also in the smallest sized fractions by GF analysis. As in the UC study, common HDL proteins like apoA-I and apoA-II were distributed across the entire HDL range. However, there were several examples of proteins that exhibited different GF elution patterns than expected from the UC data.

For example, apoL-I was exclusively found in the most dense particles by UC, but appeared in quite large HDL particles by GF. ApoE could be found distributed through all density subfractions by UC, but was focused in a rather tight pattern of larger HDL particles by GF. Although the correlation between particle density and size is not absolute, these observations lend support to the idea that high salt or sheer conditions encountered during UC separation of plasma may alter the distribution of certain

72 proteins across HDL subpopulations. Indeed, the fact that 14 new proteins were identified by GF indicates that UC may even completely remove some of the more weakly associated HDL proteins. This highlights the importance of developing alternative separation and analysis strategies for characterizing the HDL proteome.

We have previously proposed that protein clusters detected in our UC study may be indicative of distinct subsets of HDL particles which might display unique biological functions. The current study revealed several interesting observations that support this idea. First, we noticed the tight co-migration of apoL-I and haptoglobin related protein

(HGRP). These are the active protein components of the trypanosome lytic factor (TLF) which is an HDL particle shown to have specific lytic activity against Trypansoma bruceii, a protozoan parasite from the class of organisms responsible for African

Sleeping Sickness 38. These proteins co-elute in larger sized HDL fractions and also seem to co-migrate to a lesser extent in LDL sized particles. This pair comprises one of the few biochemically characterized HDL subspecies with a defined function. Second, we noted several additional co-migrating pairs including apoE and apoM, apoC-III and complement C4-B, complement C2 and insulin-like growth factor binding protein, and complement factor B and prothrombin. Although it is possible that these proteins may have co-migrated by pure coincidence, it is reasonable to propose that they may participate in potential structural interactions that sequester them to the same set of

HDL particles, perhaps to perform an as yet undefined function like the apoL-I and

HGRP pair. In addition to sharing the similar GO functions, many of these newly identified proteins interact with known HDL proteins through physical interactions.

When we examined the 47 proteins in the human protein-protein interaction network

73 from the Human Protein Reference Database (HPRD) 39, we found 8 out of the 14 new

HDL proteins have a direct interaction with the known HDL proteins, and an additional 3 can be connected to known HDL proteins by one intermediate protein (Appendix 2:

Supplemental Figure 1). This supports the possibility of potential roles for these new proteins in HDL function. Further work, using additional orthogonal separation techniques as well as immunoprecipitation experiments, will be required to test this hypothesis.

Conclusions

We have developed new separation and analysis technologies that remove practical barriers to evaluating the HDL proteome using non-centrifugal separation techniques.

This method identified new HDL associated proteins and added more weight to the growing body of evidence that distinct HDL subparticles exist that contain distinct sets of interacting proteins. Currently, therapeutic studies on cardiovascular disease are directed at raising total plasma HDL cholesterol (i.e. niacin and CETP inhibitors) without regard for the specific subspecies that may be altered. Further study of HDL subspecies may result in the identification of particles with superior cardioprotective properties or altogether new HDL functions. The methods described here will open the door for further studies of the HDL proteome using alternative non-density based lipoprotein separation strategies. A more complete knowledge of the exact protein compositions of individual subspecies and their functions may help to focus development of HDL modification strategies.

Reference List

1 Scanu AM, Wisdom C. Serum lipoproteins structure and function. Annu Rev

Biochem 1972;41:703-30.

2 Franceschini G, Maderna P, Sirtori CR. Reverse cholesterol transport: physiology

and pharmacology. Atherosclerosis 1991;88(2-3):99-107.

3 Gordon T, Castelli WP, Hjortland MC, Kannel WB, Dawber TR. High density

lipoprotein as a protective factor against coronary heart disease. The Framingham

Study. Am J Med 1977;62(5):707-14.

4 Gordon DJ, Probstfield JL, Garrison RJ, Neaton JD, Castelli WP, Knoke JD,

Jacobs DR, Jr., Bangdiwala S, Tyroler HA. High-density lipoprotein cholesterol and

cardiovascular disease. Four prospective American studies. Circulation

1989;79(1):8-15.

5 Mackness MI, Arrol S, Durrington PN. Paraoxonase prevents accumulation of

lipoperoxides in low-density lipoprotein. FEBS Lett 1991;286(1-2):152-4.

6 Cockerill GW, Rye KA, Gamble JR, Vadas MA, Barter PJ. High-density lipoproteins

inhibit cytokine-induced expression of endothelial cell adhesion molecules.

Arteriosclerosis, Thrombosis & Vascular Biology 1995;15(11):1987-94.

7 Tolle M, Pawlak A, Schuchardt M, Kawamura A, Tietge UJ, Lorkowski S, Keul P,

Assmann G, Chun J, Levkau B, van der Giet M, Nofer JR. HDL-associated

lysosphingolipids inhibit NAD(P)H oxidase-dependent monocyte chemoattractant

protein-1 production. Arterioscler Thromb Vasc Biol 2008;28(8):1542-8.

8 Davidsson P, Hulthe J, Fagerberg B, Camejo G. Proteomics of apolipoproteins and

associated proteins from plasma high-density lipoproteins. Arterioscler Thromb

Vasc Biol 2010;30(2):156-63.

9 Shih DM, Gu L, Xia YR, Navab M, Li WF, Hama S, Castellani LW, Furlong CE,

Costa LG, Fogelman AM, Lusis AJ. Mice lacking serum paraoxonase are

susceptible to organophosphate toxicity and atherosclerosis. Nature

1998;394(6690):284-7.

10 Hoff HF, O'Neil J, Chisolm GM, III, Cole TB, Quehenberger O, Esterbauer H,

Jurgens G. Modification of low density lipoprotein with 4-hydroxynonenal induces

uptake by macrophages. Arteriosclerosis 1989;9(4):538-49.

11 Shao B, Oda MN, Vaisar T, Oram JF, Heinecke JW. Pathways for oxidation of

high-density lipoprotein in human cardiovascular disease. Curr Opin Mol Ther

2006;8(3):198-205.

12 Gordon S, Durairaj A, Lu J, Davidson WS. High-Density Lipoprotein Proteomics:

Identifying New Drug Targets and Biomarkers by Understanding Functionality.

Current Cardiovascular Risk Reports 4[Volume 4, Number 1]. 1-5-2010. Current

Medicine Group LLC.

13 Vaisar T, Pennathur S, Green PS, Gharib SA, Hoofnagle AN, Cheung MC, Byun J,

Vuletic S, Kassim S, Singh P, Chea H, Knopp RH, Brunzell J, Geary R, Chait A,

Zhao XQ, Elkon K, Marcovina S, Ridker P, Oram JF, Heinecke JW. Shotgun

proteomics implicates protease inhibition and complement activation in the

antiinflammatory properties of HDL. J Clin Invest 2007;117(3):746-56.

14 van't HF, Havel RJ. Metabolism of apolipoprotein E in plasma high density

lipoproteins from normal and cholesterol-fed rats. J Biol Chem

1982;257(18):10996-1001.

15 Cham BE, Knowles BR. A solvent system for delipidation of plasma or serum

without protein precipitation. J Lipid Res 1976;17(2):176-81.

16 Keller A, Nesvizhskii AI, Kolker E, Aebersold R. Empirical statistical model to

estimate the accuracy of peptide identifications made by MS/MS and database

search. Anal Chem 2002;74(20):5383-92.

17 Nesvizhskii AI, Keller A, Kolker E, Aebersold R. A statistical model for identifying

proteins by tandem mass spectrometry. Anal Chem 2003;75(17):4646-58.

18 Davidson WS, Silva RA, Chantepie S, Lagor WR, Chapman MJ, Kontush A.

Proteomic analysis of defined HDL subpopulations reveals particle-specific protein

clusters: relevance to antioxidative function. Arterioscler Thromb Vasc Biol

2009;29(6):870-6.

19 Usui S, Hara Y, Hosaki S, Okazaki M. A new on-line dual enzymatic method for

simultaneous quantification of cholesterol and triglycerides in lipoproteins by

HPLC. J Lipid Res 2002;43(5):805-14.

20 An Trinh. LRA (Lipid Removal Agent) ...... Synthetic calcium silicate hydrate for

the selective removal of lipids, endotoxins and other bio-organic molecules. Sigma-

Aldrich Catalogue . 2010.

21 Havel RJ, EDER HA, BRAGDON JH. The distribution and chemical composition of

ultracentrifugally separated lipoproteins in human serum. J Clin Invest

1955;34(9):1345-53.

22 Collins LA, Mirza SP, Kissebah AH, Olivier M. An integrated approach for the

comprehensive characterization of lipoproteins from human plasma using FPLC

and nano-HPLC-tandem mass spectrometry. Physiol Genomics 2009.

23 Tenner AJ. Functional aspects of the C1q receptors. Behring Inst Mitt

1993;(93):241-53.

24 Matsushita M, Endo Y, Hamasaki N, Fujita T. Activation of the lectin complement

pathway by ficolins. Int Immunopharmacol 2001;1(3):359-63.

25 Cuchel M, Rader DJ. The role of high density lipoproteins in thrombosis.

ScientificWorldJournal 2002;2:89-95.

26 Heller M, Stalder D, Schlappritzi E, Hayn G, Matter U, Haeberli A. Mass

spectrometry-based analytical tools for the molecular protein characterization of

human plasma lipoproteins. Proteomics 2005;5(10):2619-30.

27 Karlsson H, Leanderson P, Tagesson C, Lindahl M. Lipoproteomics II: mapping of

proteins in high-density lipoprotein using two-dimensional gel electrophoresis and

mass spectrometry. HDL composition 2005;5(5):1431-45.

28 Han YP, Yan C, Garner WL. Proteolytic activation of matrix metalloproteinase-9 in

skin wound healing is inhibited by alpha-1-antichymotrypsin. J Invest Dermatol

2008;128(9):2334-42.

29 Plotnick MI, Rubin H, Schechter NM. The effects of reactive site location on the

inhibitory properties of the serpin alpha(1)-antichymotrypsin. J Biol Chem

2002;277(33):29927-35.

30 Becerra SP, Sagasti A, Spinella P, Notario V. Pigment epithelium-derived factor

behaves like a noninhibitory serpin. Neurotrophic activity does not require the

serpin reactive loop. J Biol Chem 1995;270(43):25992-9.

31 Houenou LJ, D'Costa AP, Li L, Turgeon VL, Enyadike C, Alberdi E, Becerra SP.

Pigment epithelium-derived factor promotes the survival and differentiation of

developing spinal motor neurons. J Comp Neurol 1999;412(3):506-14.

32 Crowe S, Wu LE, Economou C, Turpin SM, Matzaris M, Hoehn KL, Hevener AL,

James DE, Duh EJ, Watt MJ. Pigment epithelium-derived factor contributes to

insulin resistance in obesity. Cell Metab 2009;10(1):40-7.

33 Wewer UM, Ibaraki K, Schjorring P, Durkin ME, Young MF, Albrechtsen R. A

potential role for tetranectin in mineralization during osteogenesis. J Cell Biol

1994;127(6 Pt 1):1767-75.

34 Quantock AJ, Meek KM, Chakravarti S. An x-ray diffraction investigation of corneal

structure in lumican-deficient mice. Invest Ophthalmol Vis Sci 2001;42(8):1750-6.

35 Ezura Y, Chakravarti S, Oldberg A, Chervoneva I, Birk DE. Differential expression

of lumican and fibromodulin regulate collagen fibrillogenesis in developing mouse

tendons. J Cell Biol 2000;151(4):779-88.

36 Twigg SM, Kiefer MC, Zapf J, Baxter RC. Insulin-like growth factor-binding protein

5 complexes with the acid-labile subunit. Role of the carboxyl-terminal domain. J

Biol Chem 1998;273(44):28791-8.

37 Domene HM, Hwa V, Argente J, Wit JM, Camacho-Hubner C, Jasper HG, Pozo J,

van Duyvenvoorde HA, Yakar S, Fofanova-Gambetti OV, Rosenfeld RG. Human

acid-labile subunit deficiency: clinical, endocrine and metabolic consequences.

Horm Res 2009;72(3):129-41.

38 Thomson R, Samanovic M, Raper J. Activity of trypanosome lytic factor: a novel

component of innate immunity. Future Microbiol 2009;4:789-96.

39 Keshava Prasad TS, Goel R, Kandasamy K, Keerthikumar S, Kumar S,

Mathivanan S, Telikicherla D, Raju R, Shafreen B, Venugopal A, Balakrishnan L,

Marimuthu A, Banerjee S, Somanathan DS, Sebastian A, Rani S, Ray S, Harrys

Kishore CJ, Kanth S, Ahmed M, Kashyap MK, Mohmood R, Ramachandra YL,

Krishna V, Rahiman BA, Mohan S, Ranganathan P, Ramabadran S, Chaerkady R,

Pandey A. Human Protein Reference Database--2009 update. Nucleic Acids Res

2009;37(Database issue):D767-D772.

Table 2-1. Quantitative binding of ultracentrifugally isolated LDL and HDL lipids by LRA.

Figure 2-1. Elution profiles from Superose 6 (2x) and Superdex 200 (3x) size exclusion chromatography configurations. 370 μl of fresh human plasma from a normal male donor was analyzed by a tandem Superose 6 setup (a) or a triple Superdex setup (b) as described in the Experimental section. Total protein (closed circles, determined by Lowry assay) and total cholesterol (open circles, enzymatic assay) profiles across the fractions are shown. Peak designations refer to 1: VLDL, 2: LDL, 3: HDL, 4: lipid-free plasma proteins.

Figure 2-2. SDS PAGE comparison of total HDL preparations derived from ultracentrifugation (UC) and gel filtration (GF) chromatography. A 4-15% SDS-PAGE analysis of total HDL isolated by UC (lane 1) or pooled HDL fractions from the triple Superdex 200 gel filtration separation (lane 2) is shown. The gel was stained with coomassie blue.

Figure 2-3. Ability of calcium silicate hydrate (LRA) to bind ultracentrifugally isolated human plasma lipoproteins. UC isolated LDL (panel a) or HDL (panel b) were analyzed by SDS PAGE prior to incubation with LRA (lane 1 of each panel). The resulting flow through is shown in lane 2 and the proteins retained on the resin after boiling with SDS sample buffer is shown in lane 3 of each panel. The gels were stained with coomassie blue. Panel c shows the same experiment in panel b, except that apoA-I was detected by western blot using an anti-human apoA-I antibody.

Figure 2-4. Ability of LRA to bind phospholipid-containing particles from fractions collected by gel filtration chromatography. a) Two identical samples of human plasma were fractionated on the triple Superdex gel filtration set up. One set of fractions was incubated with LRA under the conditions described in the Experimental section (closed circles) while the other fraction was left untreated (open circles). The traces show the phospholipid content of each fraction as determined by enzymatic assay. b) Triple Superdex gel filtration fraction 23 (lane 1) was incubated with LRA for 30 min and then the supernatant containing unbound components was removed (lane 2). The LRA was washed with buffer and bound proteins were recovered from LRA by boiling in SDS-sample buffer (lane 3). SDS-PAGE was carried out on a 4-15% gel and stained with coomassie brilliant blue. c) ApoB containing lipoproteins (fraction 16) were analyzed in the same manner as described for panel b.

Figure 2-5. Examples of elution profile shifts for proteins upon ether delipidation of fresh human plasma. Protein distribution profiles of selected proteins from untreated (open circles) and ether delipidated plasma (closed circles) after separation by the triple Superdex setup. The distribution of each protein is represented as spectral count per fraction measured by ESI-MS/MS. Panel (a) shows plasminogen, which fails to exhibit a molecular size shift in response to ether delipidation and therefore is not associated with lipid, and panel (b) shows complement C3 which does shift, indicating an association with lipid. Representative data shown from two independent experiments are shown.

Figure 2-6. Lipid-associated proteins identified in the plasmas of 3 normolipidemic donors. The proteins included in this list met all identification criteria laid out in the Experimental section and showed a shift in elution pattern after ether delipidation - indicating lipid association. The proteins are arranged according to the sum of identified peptides across all gel filtration fractions for all 3 subjects. Proteins indicated with an asterisk have not been previously described as lipid associated proteins, to our knowledge.

Figure 2-7. Distribution patterns of common HDL associated proteins across gel filtration fractions. For each protein, the number of spectral counts in a given fraction is represented by bar height. The values represent the sum of counts from 3 subjects.

Figure 2-8. Triple Superdex distribution profiles for identified lipid-associated proteins. For each fraction, the relative abundance (determined by peptide count) is shown. A value of 1.0 was assigned to the fraction containing the highest peptide count for that particular protein and all other fractions were scaled from there. The relative abundance of each can also be assessed by the color of the square with blue representing 0 detected peptides and red representing the highest number.

Figure 2-9. Gene Ontology functional associations of newly identified lipoprotein associated proteins. Identified proteins are grouped by functional category (left column) and enrichment of a particular function for either newly identified or previously established proteins is presented as the number of proteins possessing function divided by total number of proteins in group. P values are given in parentheses.

Chapter 3. A novel correlation analysis to determine HDL subspecies protein composition

Introduction

The functional diversity of HDL is likely a result of its complex compositional heterogeneity. As mentioned in previous chapters, recent studies have taken advantage of high-resolution mass spectrometry techniques to identify about 55 distinct protein components associated with HDL particles 1. Nearly all of these studies used ultracentrifugation to isolate HDL. Here we developed alternative separation techniques for lipoprotein isolation and chose not to include HDL isolated by ultracentrifugation because we know that this technique can alter the composition and functionality of HDL due to long term exposure to high g forces and high salt concentrations 2. In Chapter 2, we used gel filtration chromatography to collect plasma lipoproteins in 17 fractions based on size 3. We then used electrospray ionization mass spectrometry (ESI-MS/MS) to analyze the distribution of proteins across size fractions containing phospholipid.

This study showed that lipid associated proteins displayed a differential distribution across size fractions such that each fraction had its own unique complement of lipid associated proteins and supports our global hypothesis that the total pool of HDL in an individual is composed of numerous proteomically distinct subpopulations.

One major advancement of our previous study was the use of LRA to isolate PL associated proteins from lipid free proteins which co-fractionate with lipoproteins.

Because this separation technique is based on size rather than density, the lipoprotein components are not totally isolated from lipid free proteins which often co-fractionate.

Without first isolating the lipoprotein associated proteins it would not have been possible to know whether the proteins identified by mass spectrometry were in fact associated with a lipidated particle. The use of the separation techniques introduced in this chapter was made possible by this advance as they are also based on physical properties other than density and therefore are likely to encounter the same issue with co-migrating lipid free proteins.

A limitation of our previous study was that we could not obtain high enough resolution with the gel filtration chromatography to isolate a single subspecies of HDL in a single fraction, preventing us from determining the exact protein composition of HDL particles. Surely, it would be extremely difficult to fractionate single HDL species when separating by any available separation technique. To address this issue and attempt to determine the protein compositions of distinct HDL subspecies we have developed a multi-faceted correlation analysis which combines proteomics data from multiple plasma separation techniques, allowing us to identify lipid associated protein components which consistently co-fractionate when separated by various physicochemical properties

Figure 3-1. Clusters of lipid associated proteins which tended to co-migrate consistently across different separation techniques were assigned higher correlation scores. High scoring protein clusters were identified as potential components of a single HDL particle and were confirmed using cross-linking studies.

Experimental

(0.105 M) as an anticoagulant. Cellular components were pelleted by centrifugation at

~1,590 x g for 15 minutes in a Horizon mini-E (Quest Diagnostics) at room temperature.

Plasma was stored at 4ºC until applied to a separation technique, always within 16 hours. Samples were never frozen.

Plasma separation techniques:

Gel filtration chromatography. Plasma (370 µL) was applied to three Superdex 200

10/300 GL columns (GE Healthcare) arranged in series as described in Chapter 2 3.

Anion exchange chromatography. Plasma (370 µL) was mixed with buffer A (50 mM

Tris, 1 mM EDTA) for a total volume of 2 mL and applied to a MonoQ 5/50 GL anion exchange column (GE Healthcare) at a flow rate of 1 mL/min. A gradient of buffer B

(Buffer A + 500 mM sodium perchlorate) was used to elute plasma components from the resin and 1 mL fractions were collected. This method was adapted from 4.

Isoelectric focusing. Plasma (3 mL) was prepared for separation by mixing with 56 mL of water and 3 mL of Bio-lyte 5/7 ampholyte solution (BioRad). Mixture was loaded

93 into Rotofor system (BioRad) and run at constant 15 W until voltage stabilized (approx.

4 hours). Fractions were collected using supplied fraction collector.

Collected fractions from each plasma separation technique were analyzed for phospholipid (PL, enzymatic kit from Wako), cholesterol (CH, enzymatic kit from Pointe

Scientific) and total protein content (Bradford assay).

Isolation of lipid associated proteins and mass spectrometry. Fractions collected from each technique were applied to a calcium silica hydrate resin which binds to phospholipid (LRA, Supelco) to isolate only those proteins which are associated with phospholipid. We have previously demonstrated the binding of lipids and HDL proteins to LRA in Chapter 2 3.Lipid associated proteins were identified using electrospray ionization mass spectrometry. All mass spectrometry procedures and analyses were performed exactly as described in Chapter 2.

Correlation analysis. The strategy for our correlation analysis is outlined in Figure 3-1.

Protein distribution profiles for each separation technique were generated using the mass spectrometry data. With the distribution data, we used Pearson correlation to calculate the similarity between the abundance profiles for all combinations of protein

th pairs (Eq. 1), where Xi and Yi are the abundance value of protein X and Y in the i fraction. Higher scores indicate protein pairs that were more consistently collected in the same fraction by each individual separation technique. The correlation scores from each of the three separation techniques for each pair of proteins were summed to

94 obtain a combined correlation score. This score was used to determine the likelihood of particle co-habitance by the protein pair.

()()XXYY i ii rXY,  (Eq.1), ()()XX22 YY iiii

Results

Three plasma separation techniques. Each plasma separation technique was optimized to produce optimum separation of lipid containing particles across collected fractions. Figure 3-2 displays the distribution profiles of protein, phospholipid and cholesterol across fractions collected by gel filtration chromatography. This technique was described in detail in Chapter 2.

Plasma separation by anion exchange chromatography produced very different lipid and total protein distribution patterns. There was an initial protein peak across fractions 1-6 with very little lipid. This peak likely contains lipid-free plasma proteins with either positive or neutral charge as they were not retained on the column resin.

Lipids were distributed across the remaining fractions 7 through 25. There appear to be three distinct lipid peaks from left to right. The first (Fractions 7-11) is cholesterol rich with small amounts of phospholipid and a relatively large amount of protein. The second (Fractions 12-19) is phospholipid rich with lower amounts of cholesterol and a relatively moderate amount of protein. And the third (Fractions 20-25) is cholesterol rich with a small amount of protein. From this data it is difficult to distinguish between

95 lipoprotein classes i.e. we cannot determine if separation occurred between HDL or

LDL, for example.

Separation of plasma by anion exchange again produced entirely different distribution patterns from either of the previous techniques. Here lipids are distributed across 15 fractions (Fractions 6-20) as two peaks. The first has approximately equal content of phospholipid and cholesterol while the second has almost a 2 to 1 ratio of phospholipid to cholesterol. Again because we are separating using techniques other than density we cannot distinguish lipoprotein classes, it is possible that each of these lipid peaks represent a mixture high and low density lipoproteins.

As anticipated, when plasma was separated by different physical properties, the distribution profiles of these components varied such that each separation technique generated unique distribution profiles. In the next section we take advantage of the varied lipoprotein distribution patterns in these techniques as it provides an optimal opportunity for the tracking of lipoprotein protein components to determine protein to protein association on single HDL particles.

Proteomics. Phospholipid associated proteins in each fraction were isolated using

Lipid Removal Agent resin (Supelco) and analyzed by mass spectrometry. Analyzing each fraction independently allowed us to generate distribution profiles for proteins across fractions for each plasma separation technique. This distribution profile data is best represented with heat maps. A heat map showing lipid associated protein distribution by gel filtration chromatography can be found in Chapter 2 and those for anion exchange and isoelectric focusing are included here (Figures 3-5 and 3-6). From

96 this data we can better track the distributions of the lipoprotein classes. By tracking the distribution of apoB and apoA-I we can better tell where the VLDL and LDL or HDL have migrated by these techniques. By gel filtration chromatography, size often roughly relates to density resulting a separation pattern where the LDL and VLDL are collected as one single peak in the early eluting fractions 14-18 and the HDL is collected as a broad peak that stretches from fractions 20-30 (described in Chapter 2). In anion exchange chromatography, apoB appears as two peaks across fractions, a small amount in 9-11 and the majority in 18-26 while apoA-I is found in all phospholipid containing fractions and the remaining proteins are scattered across fractions. Because of this it appears that we do not get separation of VLDL, LDL and HDL based on charge. By isoelectric focusing, apoB is found primarily in fractions 4-10 with very little relative amounts found elsewhere. ApoA-I is again found in all phospholipid containing fractions while the remaining proteins are distributed differently across fractions.

Correlation analysis. The goal of our correlation analysis was to identify pairs of proteins which are likely to reside on a single phospholipid particle. The analysis uses a scoring system to identify pairs of proteins whose distribution profiles are similar within a given separation technique. Proteins which are consistently identified in the same fraction are given higher correlation scores. The power of this strategy lies in the use of multiple separation techniques. The individual correlation scores for a given pair of proteins for each separation technique are combined to produce an overall correlation score that is representative of the likelihood that those proteins reside together on a single particle. Table 3-1 contains the top 100 ranked results of our correlation analysis

97 representing protein pairs scoring in the top 15%. In total, scores were assigned to 741 protein pairs. It was no surprise that the top two most abundant HDL proteins apoA-I

(~65% protein by mass) and apoA-II (~15% protein by mass) received the highest combined correlation score as they are found together in most fractions and biochemical studies have demonstrated that they reside together on a significant portion of HDL particles 5. Upon manual examination, we find a high degree of overlap in the distribution profiles of high ranking protein pairs, Figure 3-3 A shows as an example the distribution patterns by each separation technique for inter-alpha-trypsin inhibitor H4

(ITIH4) and insulin like growth factor binding protein (ALS) of the second ranking protein pair from Table 3-1. In these traces it is easy to see the high degree of overlap in the distribution of these two proteins. One of only a few biochemically confirmed HDL subspecies is known as Trypanosome lytic factor (TLF, described in Chapter 1). This particle is known to contain apoA-I as well as two additional proteins, apoL-I and haptoglobin related protein (HPTR). To validate our approach we manually examined the distribution profiles of these proteins already known to exist together on a single

HDL particle. Figure 3-3 B displays the distribution patterns for apoL-I and HPTR across fractions for each of the separations techniques. The high degree of overlap in the profiles of these two proteins suggests that our approach can successfully identify a previously biochemically determined HDL subspecies. This established pair scored within the top 5% of protein pairs at 34th (Table 3-1) providing further biochemical support for our correlation approach. The remaining protein pairs listed in table 3-1 represent potentially novel protein interactions on HDL, the majority are not yet reported to our knowledge.

Discussion

In this study we have expanded on our previous work with the goal of identifying the protein components of single subspecies of HDL. Recent applications of modern mass spectrometry technologies to analyzing the proteome of HDL suggest that the protein composition of HDL is much more expansive than previously believed. Although these

“minor” HDL proteins may only make up about 20% of the total protein content of HDL by mass, it is possible that these proteins may be the components of subspecies of HDL conferring significant function even at low particle number. One example of such an

HDL subspecies is already known to exist, TLF (described above) has potent trypanolytic activity while making up only a very small fraction of total HDL 6.

The results of our correlation analysis resulted in the assignment of correlation scores to 741 protein pairs with the top 15% listed in Table 3-1. This study represents the first step toward a more global characterization of HDL subspecies. The significance and functional relevance of the protein pairs identified here have not yet been examined and invite further investigation. A full listing of identified protein pairs and their correlation scores can be found in Appendix 3. For now, we can only speculate about potential interactions between these proteins or ways in which their presence together on a single particle may influence the known functionality of either protein or result in a new previously unsuspected functional attribute. For some of the identified protein pairs this speculation is easier than for others. For example, ranked

3rd in Table 3-1 is the pairing of complement components C1s and C4b. C1s combines

99 with other subcomponents to form C1, a serine protease responsible for the cleavage and subsequent activation of various downstream components in the “classical pathway” of the complement cascade 7. One immediate downstream target is C4, cleavage of which produces the C4b fragment which itself has further downstream effects in the complement cascade. Based on this information it is not too difficult to develop hypotheses for why these two components of a common pathway of the complement cascade may find themselves sitting together on an HDL particle. Perhaps the HDL particle provides a scaffold for these proteins to find each other thus facilitating the cleavage event or alternatively there association here may be only remnants of this cleavage event but their continued physical association may be of yet undiscovered functional importance. The very presence of these proteins on HDL suggests that HDL may be playing a role in the early activation stages of the classical complement pathway. It is likely that many more examples like this exist among the top ranking protein pairs in Table 3-1.

This is the first use of these specific separation techniques for the proteomic analysis of lipoproteins, to our knowledge. Proteomic analysis of lipoproteins separated by physical properties aside from density was made possible by the use of the LRA resin described in Chapter 2. One limitation of the anion exchange and isoelectric focusing separation techniques used here is that they do not provide distinct separation between HDL and lower density lipoproteins as can be achieved with gel filtration or ultracentrifugation. This brings up an interesting point in the classification of lipoproteins, we currently use a definition based on density because of their original isolation by ultracentrifugation. This has stuck because this type of separation, along

100 with the less used gel filtration methods, was the best way to isolate lipoproteins from lipid free plasma proteins. But with the advance of the LRA technique we can now use alternative separation strategies that result in the same pool of lipoproteins spread across a different spectrum that is based on other physical properties and although this might result in overlap between traditionally defined HDL and LDL is still useful.

Because of this it is possible that some of the protein pair identifications here could represent those on VLDL or LDL. To compensate for this we limited our correlation analysis to include only those proteins identified as being “likely” to be associated with

HDL according to Table 1-1, meaning they have been found in association with total

HDL in at least 3 proteomics studies of HDL. Another limitation of the present study is the restriction of this analysis to pairs of proteins. A useful extension of this study would be to explore interaction networks containing three or more proteins, as it is likely that many HDL particles contain more than just 2 proteins. The limitations of available technology are also at play. In complex protein samples studied here we are restricted by the limits of detection of the mass spectrometry equipment used. Many of the proteins we are searching for comprise only a small percentage of the total protein content of our samples, some less than 1%. It can be difficult for an instrument to identify peptides for common low abundance HDL proteins, such as the important enzymes CETP or lecithin cholesterol acyl transferase (LCAT), among such high background of peptides from the most abundant proteins. As more sensitive technologies become available even more in-depth analyses will become possible.

The study of protein interactions on subspecies of HDL has vast potential for the further understanding of HDL’s diverse functionality but also for the discovery of novel

101 functions not yet attributed to this dynamic class of lipoproteins. Further analysis of these protein components and the ways in which they interact on HDL will provide useful information for the further development of HDL as a biomarker and potential therapeutic target for multiple disease states.

102

Reference List

1 Gordon S, Durairaj A, Lu J, Davidson WS. High-density lipoprotein proteomics:

Identifying new drug targets and biomarkers by understanding functionality. Curr

Cardio Risk Rep 2010;4(Volume 4, Number 1):1-8.

2 van't HF, Havel RJ. Metabolism of apolipoprotein E in plasma high density

lipoproteins from normal and cholesterol-fed rats. J Biol Chem 1982;257(18):10996-

1001.

3 Gordon SM, Deng J, Lu LJ, Davidson WS. Proteomic characterization of human

plasma high density lipoprotein fractionated by gel filtration chromatography. J

Proteome Res 2010;9:5239-49.

4 Hirowatari Y, Yoshida H, Kurosawa H, Doumitu KI, Tada N. Measurement of

cholesterol of major serum lipoprotein classes by anion-exchange HPLC with

perchlorate ion-containing eluent. J Lipid Res 2003;44(7):1404-12.

5 Cheung MC, Albers JJ. Characterization of lipoprotein particles isolated by

immunoaffinity chromatography. Particles containing A-I and A-II and particles

containing A-I but no A-II. J Biol Chem 1984;259(19):12201-9.

6 Raper J, Fung R, Ghiso J, Nussenzweig V, Tomlinson S. Characterization of a novel

trypanosome lytic factor from human serum. Infect Immun 1999;67(4):1910-6.

103

7 Schroeder H, Skelly PJ, Zipfel PF, Losson B, Vanderplasschen A. Subversion of

complement by hematophagous parasites. Dev Comp Immunol 2009;33(1):5-13.

104

Plasma

Iso-electric Gel Filtration Anion Exchange Focusing

HDL sub- HDL sub- HDL subfractions fractions fractions

Mass Spectrometry Analysis

Protein Protein Protein distribution distribution distribution profile profile profile

Correlation Networks

Statistically Combined Network

Candidate Interactions

Figure 3-1. Correlation strategy for the identification of phospholipid particle subspecies.

105

3.0 0.10 2.5 0.08 2.0 0.06 1.5 0.04

(ug/ul) 1.0 Protein 0.02 0.5 0.0 0.00 13 15 17 19 21 23 25 27 29

Figure 3-2. Lipid and protein distribution profiles produced by gel filtration chromatography.

106

3.0 0.10 2.5 0.08 2.0 0.06 1.5 0.04

(ug/ul) 1.0 Protein 0.02 0.5 0.0 0.00 1 3 5 7 9111315171921232527

Figure 3-3. Lipid and protein distribution profiles produced by anion exchange chromatography.

107

3.0 0.06 2.5 0.05 2.0 0.04 1.5 0.03 (ug/ul)

Protein 1.0 0.02 0.5 0.01 0.0 0.00 2 4 6 8 10 12 14 16 18 20

Figure 3-4. Lipid and protein distribution profiles produced by isoelectric focusing.

108

C4b-binding protein alpha chain Serum paraoxonase/arylesterase 1 Apolipoprotein B-100 Complement C1s subcomponent Complement C4-B Apolipoprotein F Apolipoprotein E Apolipoprotein A-II Apolipoprotein A-I Apolipoprotein M Apolipoprotein C-II Apolipoprotein D Complement C3 Apolipoprotein C-III Transthyretin precursor Plasma protease C1 inhibitor protein Haptoglobin-related Apolipoprotein-L1 Inter-alpha-trypsin inhibitor heavy chain H4 Alpha-1-antichymotrypsin Pigment epithelium-derived factor Insulin-like growth factor-binding protein Clusterin (apoJ) Antithrombin-III Angiotensinogen Alpha-2-antiplasmin Alpha-2-HS-glycoprotein Alpha-1-antitrypsin Lipopolysaccharide-binding protein Complement factor B Apolipoprotein A-IV Apolipoprotein C-I Serum albumin Serum amyloid A-4 Serotransferrin Beta-2-glycoprotein 1 Protein name

Figure 3-5. Heat map displaying protein distribution patterns by anion exchange chromatography.

0.0357 0.2881 0.0847 0.0313 0.0667 0.0909 0.0441 0.1111 0.2044 0.1429 0.075 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 5 0.1786 0.5847 0.4545 0.2857 0.1429 0.3175 0.2794 0.4444 0.7127 0.7405 0.025 0.2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 6 0.2143 0.6356 0.2727 0.2857 0.2857 0.4921 0.5294 0.8571 0.2824 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 7 0.7712 0.0667 0.0909 0.1667 0.2857 0.9448 0.5714 0.0076 0.15 0.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 8 0.3485 0.3571 0.7797 0.2105 0.0833 0.1375 0.3043 0.9091 0.4286 0.6032 0.7647 0.6667 0.6519 0.2857 0.25 0.6 0.2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 9 0.3485 0.3571 0.5789 0.1695 0.6667 0.6522 0.3333 0.7143 0.1429 0.0794 0.4706 0.4444 0.1429 0.025 0.661 0.625 0.453 0.55 0.6 0.4 0.2 10 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0.0758 0.2143 0.4322 0.1667 0.0033 0.2222 0.3333 0.6842 0.7125 0.5254 0.2727 0.2778 0.5714 0.1429 0.2941 0.2222 0.2486 0.1429 0.005 0.025 0.4 11 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0.1667 0.1111 0.8889 0.2105 0.7143 0.7826 0.7188 0.6667 0.2727 0.1667 0.4286 0.2353 0.1111 0.2155 0.025 0.5 0.5 0.9 12 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0.0199 0.2857 0.4915 0.6667 0.0131 0.4444 0.4667 0.1625 0.0952 0.8696 0.2188 0.0909 0.1111 0.4286 0.1324 0.1111 0.2376 0.339 0.05 0.4 0.4 0.6 13 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0.0398 0.7881 0.6667 0.3967 0.6667 0.2222 0.8421 0.6667 0.0847 0.5652 0.2667 0.1429 0.0294 0.6667 0.1436 0.1429 0.075 0.05 0.5 0.8 0.8 0.8 14 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Fraction number Fraction 0.0199 0.3929 0.8559 0.9508 0.1111 0.3684 0.4167 0.0125 0.6957 0.0294 0.3333 0.0663 0.1429 0.175 0.8 0.5 0.4 15 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0.3929 0.8475 0.5333 0.4211 0.3333 0.5217 0.0588 0.5556 0.0773 0.1429 0.005 0.375 16 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.0075 0.5357 0.6426 0.8889 0.3333 0.1579 0.5217 0.0882 0.5556 0.0608 0.1429 0.01 0.8 0.8 17 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.0225 0.0348 0.6695 0.3333 0.2426 0.6667 0.0667 0.5652 0.0147 0.2222 0.0221 0.25 0.8 0.4 18 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.0413 0.0152 0.1875 0.2857 0.4237 0.0328 0.4444 0.3043 0.0147 0.2222 0.0055 0.393 0.425 0.75 0.4 0.5 0.4 19 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.0188 0.2576 0.9375 0.1071 0.2797 0.1667 0.0033 0.5556 0.2609 0.0055 0.2 0.2 0.2 20 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.0341 0.7273 0.5025 0.0714 0.2458 0.2222 0.2609 0.175 0.011 0.03 0.2 21 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.0375 0.2786 0.1071 0.3136 0.1667 0.0295 0.3333 0.1111 0.0833 0.3043 0.1111 0.875 0.011 0.2 0.2 0.2 22 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.0682 0.0225 0.7121 0.1875 0.2736 0.0714 0.2627 0.1667 0.0754 0.2222 0.1333 0.1739 0.0055 0.275 0.2 23 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.1705 0.0338 0.4848 0.0647 0.1111 0.2609 0.1111 0.275 0.161 0.011 24 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.1818 0.0945 0.0339 0.1667 0.2222 0.4348 0.1111 0.0055 0.375 0.03 0.2 25 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.7273 0.0404 0.0531 0.2963 0.4638 0.113 0.1 26 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.7273 0.0464 0.4058 0.065 0.113 27 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.0202 0.0332 0.0226 0.4058 0.05 28 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

109

Fraction number Protein name 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Serotransferrin 0.15 0.375 0.675 1 0.45 0.475 0.175 0.25 0.225 0.025 0 0 0 0 0 0 0 0 0 0.35 C4b-binding protein alpha chain 0 0 0.1 0 1 0 0.1 0.3 0.4 0 0 0 0 0 0 0 0 0 0 0 Complement C1r subcomponent 0 0 0 0 1 0 0 0.3333 0 0 0 0 0 0 0 0 0 0 0 0 Complement C1s subcomponent 0 0 0.25 0 1 0 0 0.75 0.5 0 0 0 0 0 0 0.5 0 0 0 0 Complement C3 0.5 0.1111 0.5926 0.0556 1 0.0741 0 0.1852 0.2407 0 0.1481 0 0 0.0185 0.0185 0.4074 0.3333 0.0741 0.0185 0.1852 Complement C4-B 0.1905 0.0476 0.1905 0 1 0.2381 0.0476 0.8095 0.7619 0.1905 0.0952 0.0476 0.0952 0 0.0952 0.2381 0.1905 0.1905 0 0 Fibrinogen alpha chain 0.3696 0.4348 0.6957 0.9565 1 0.7391 0.7609 0.8696 0.4348 0.3261 0.2609 0.2609 0.2609 0.2826 0.2391 0.3696 0.3043 0.0652 0 0 Fibrinogen gamma chain 0.3871 0.1935 0.7742 0.9355 1 0.9032 0.9032 0.8387 0.7419 0.3548 0.4516 0.3871 0.3226 0.2258 0.4194 0.6129 0.4839 0.1613 0 0.0645 Fibronectin 0.0385 0 0.2692 0 1 0.0385 0 0.5385 0.3462 0 0.0385 0 0.0385 0 0 0.2308 0 0 0 0 Inter-alpha-trypsin inhibitor heavy chain H1 0 0 0.8889 0.2222 1 0.3333 0 0 0.2222 0.1111 0.3333 0.1111 0.1111 0.2222 0.7778 0.8889 0.2222 0 0 0 Plasminogen 0.1818 0 0.3636 0 1 0.0909 0 0.9091 0.3636 0 0 0 0 0 0 0.1818 0.0909 0 0 0 Beta-2-glycoprotein 1 0.4167 0.5833 0.75 0.5 0.8333 0.5833 1 0.9167 0.6667 0.6667 0.9167 0.75 0.4167 0.5833 0.5 0.4167 0.25 0.25 0 0 Apolipoprotein B-100 0 0 0 0.0392 0.4412 0.6176 0.4706 1 0.5098 0.049 0 0.0098 0 0 0 0.0392 0.0098 0 0 0 Complement C5 0 0 0 0 1 0 0 1 0.5 0 0 0 0 0 0 0 0 0 0 0 Complement component C6 0 0 0 0 1 0.1667 0.6667 1 0 0 0 0 0 0 0 0 0 0 0 0 Complement component C8 beta chain 0 0 0 0 0.5 0 0.25 1 0.5 0.25 0 0.125 0.125 0 0 0 0.125 0 0 0 Fibrinogen beta chain 0.4872 0.3077 0.7436 0.8462 0.8718 0.8974 0.8462 1 0.7179 0.4872 0.2821 0.5128 0.5128 0.2564 0.4872 0.6923 0.4103 0.1282 0.0256 0.1026 Complement component C7 0 0 0 0 0 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0 Complement component C8 alpha chain 0 0 0 0 0 0 0 1 1 0 0 0 0.5 0 0 0 0.5 0 0 0 Apolipoprotein C-I 0.2857 0 0.1429 0.4286 0.1429 0.5714 0.5714 0.8571 1 0.2857 0.5714 0.5714 0.2857 0.1429 0 0 0 0 0 0 Complement factor H 0 0 0.1765 0 0.9412 0 0 0.9412 1 0 0.0588 0 0 0 0 0.2353 0.0588 0 0 0 Gelsolin 0.0714 0.0952 0.1667 0.1429 0.119 0.4524 0.4048 0.5952 1 0.7619 0.3333 0.1905 0.0238 0 0 0 0 0 0 0 Serum amyloid A protein 0 0 0 0 0 0 0 0 0 0.25 1 0.75 0.25 0.25 0.5 0 0 0 0 0 Tetranectin 0 0 0 0 0 0.2 0.1333 0.3333 0.6667 0.9333 1 0.6667 0.1333 0 0 0 0 0 0 0 Hemopexin 0 0 0.1364 0.1364 0.1818 0.3182 0.0909 0.5909 0.7273 0.7273 1 1 0.7727 0.7727 0.5909 0.3182 0.4091 0.3182 0 0.1818 Apolipoprotein A-I 0.0748 0.1121 0.1402 0.2523 0.1776 0.1776 0.1121 0.1308 0.1869 0.2991 0.6636 1 0.9439 0.8785 0.7383 0.5888 0.4206 0.2523 0.1308 0.1682 Apolipoprotein E 0.0303 0 0 0.0606 0.0606 0.2727 0.1515 0.4545 0.5455 0.6061 0.9091 1 0.9091 0.697 0.4242 0.2424 0.0606 0 0 0 Apolipoprotein-L1 0 0 0 0 0 0 0 0 0 0.0833 0.5 1 0.8333 0.5 0.3333 0.1667 0 0 0 0 Complement C2 0 0 0 0 0 0 0 0 0 0 0.5 1 0.5556 0.1667 0.2222 0 0.3889 0.0556 0 0 Pigment epithelium-derived factor 0 0 0 0 0.4706 0.1176 0.2353 0.4706 0.4118 0.5882 0.8824 1 0.5294 0.5294 0.6471 0.3529 0.2353 0 0 0 Serum amyloid A-4 protein 0 0 0 0 0 0 0 0 0 0.1429 0.4286 1 0.8571 0.5714 0.2857 0.1429 0.1429 0 0 0 Complement factor I 0 0 0 0 0 0 0 0 0 0.2 0.6 0.5 1 0.7 0.2 0 0.1 0 0 0 Apolipoprotein A-II 0 0.0625 0.0625 0.25 0.125 0.0625 0 0 0 0.0625 0.75 0.875 0.75 1 0.875 0.6875 0.625 0.125 0.0625 0 Apolipoprotein A-IV 0.0286 0.0286 0.0714 0.0857 0.0143 0.0714 0.0143 0.0571 0.1 0.1571 0.3429 0.6714 0.8429 1 0.6571 0.5286 0.3 0.0286 0 0 Apolipoprotein C-II 0 0 0 0 0 0 0 0 0 0 0.1667 0.3333 0.8333 1 0.6667 0.8333 0.1667 0.1667 0 0 Insulin-like growth factor-binding protein ALS 0 0 0 0 0 0 0 0 0 0 0.6667 0.5 0.6667 1 0.8333 0.5 0.5 0.1667 0 0 Inter-alpha-trypsin inhibitor heavy chain H2 0 0 0.3333 0 0.75 0.0833 0 0.0833 0.0833 0.0833 0.1667 0.4167 0.5 1 0.6667 0.4167 0.0833 0 0 0 Kallistatin 0 0 0 0 0 0 0 0 0 0 0.375 0.75 0.875 1 0.375 0.125 0.25 0.125 0 0 Serum paraoxonase/arylesterase 1 0 0 0 0 0 0 0 0 0 0 0 0 0.125 1 0.75 0.625 0.375 0 0 0 Apolipoprotein M 0 0 0 0 0 0 0.125 0 0 0 0.125 0.625 0.375 0.875 1 0.625 0.25 0 0 0 Heparin cofactor 2 0 0 0.15 0 0.1 0.1 0 0 0 0 0.05 0.35 0.75 0.85 1 0.9 0.65 0.15 0 0 Complement component C9 0 0 0.0833 0.1667 0.3333 0.25 0.0833 0.0833 0 0.0417 0.2917 0.2917 0.375 0.8333 1 0.6667 0.4167 0.0833 0 0 Protein Z-dependent protease inhibitor 0 0 0 0 0 0 0 0 0 0 0 0.25 0.25 0.75 1 0.75 0.5 0 0 0 Antithrombin-III 0 0 0 0.0278 0.0278 0 0 0 0 0 0 0.0833 0.5 0.8611 1 0.5833 0.3333 0.0278 0 0 Apolipoprotein D 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0.5 0 0 0 Apolipoprotein F 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.3333 1 0.3333 0 0 0 Haptoglobin-related protein 0 0 0 0 0 0 0 0 0 0 0 0 0.5909 0.5909 0.6818 1 0 0 0 0 Prothrombin 0 0 0.2308 0.2308 0.8462 0.1538 0 0.2308 0.3077 0 0.1538 0.3846 0.3077 0.6923 0.4615 1 0.4615 0.1538 0.0769 0.1538 Vitamin D-binding protein 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0.1667 0.1667 0 0 Alpha-1-antitrypsin 0.1184 0.0526 0.1184 0.1447 0.1316 0.0921 0.0789 0.0658 0.0395 0.0395 0.0658 0.1447 0.2105 0.4211 0.5789 0.8289 1 0.3816 0.3684 0.3684 Alpha-2-antiplasmin 0 0 0 0 0 0 0 0 0 0 0 0 0.1818 0.9091 0.6364 0.7273 1 0.2727 0 0 Alpha-2-macroglobulin 0.0357 0.1429 0.5714 0 0.7857 0 0 0 0 0 0 0.0357 0.0714 0.1786 0.2143 0.7143 1 0.0714 0.0714 0.4286 Complement factor B 0 0 0.0909 0.0909 0.3333 0.3636 0.1515 0.1212 0.1818 0.4545 0.697 0.9697 0.6061 0.697 0.8788 0.7576 1 0.7576 0.1212 0.0303 Haptoglobin 0.0465 0.1395 0.2093 0.2093 0.2093 0.2093 0.093 0.0233 0 0 0.0698 0.2326 0.3721 0.3953 0.5814 0.7907 1 0.814 0.9767 0.7442 Inter-alpha-trypsin inhibitor heavy chain H4 0 0.0357 0.1429 0.1429 0.2143 0.1786 0 0 0 0 0.1429 0.3571 0.4643 0.7857 0.8571 0.9286 1 0.5357 0.25 0.1071 Vitronectin 0 0 0.2 0.3 0.4 0.1 0 0.1 0.1 0 0 0 0 0.4 0.4 1 1 0.6 0.5 0.7 Plasma protease C1 inhibitor 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.5 1 0.25 0 0.25 Transthyretin 0.0476 0.0476 0.1905 0.1429 0.2381 0.1905 0.0476 0 0 0 0.1905 0.5238 0.4286 0.5714 0.619 0.7619 1 0.7143 0.4286 0.1429 Histidine-rich glycoprotein 0 0 0 0 0.7143 0.1429 0 0.2857 0.8571 0.2857 0.7143 0.7143 0.8571 0.5714 0.7143 0.7143 1 1 0 0 Apolipoprotein C-III 0 0 0 0.1429 0.1429 0.1429 0.2857 0.2857 0 0 0.2857 0.4286 0.4286 0.7143 0.5714 0.4286 0.7143 1 0.2857 0.2857 Clusterin 0 0 0.1 0.15 0.1 0.1 0.1 0 0 0 0 0.05 0 0.15 0.25 0.8 0.8 1 0.8 0.7 Lipopolysaccharide-binding protein 0 0 0 0 0 0 0 0 0 0 0.5 0 0.5 0 0 0 0.5 1 0 0 Kininogen-1 0 0 0.1111 0.1111 0.1111 0 0 0 0 0 0 0.1111 0 0 0 0 0.5556 0.4444 1 0.7778 Alpha-2-HS-glycoprotein 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.125 0.625 0.125 0.75 1 Lumican 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.3333 1 Alpha-1-antichymotrypsin 0.0323 0 0.1613 0 0.0968 0 0 0 0 0 0 0 0 0.0323 0.1613 0.3548 0.6452 0.4839 0.6129 1

Figure 3-6. Heat map displaying protein distribution patterns by isoelectric focusing.

110

Table 3-1. Results of correlation analysis. Correlation Score Rank Protein A Protein B GF AE IEF Combined 1 APOA1 APOA2 0.89 0.80 0.86 2.55 2 ALS ITIH4 0.87 0.79 0.49 2.15 3 C1S CO4B 0.41 0.91 0.78 2.09 4 CLUS ITIH4 0.61 0.80 0.55 1.96 5 A1BG A2AP 0.82 0.83 0.27 1.92 6 APOH HEMO 0.86 0.71 0.31 1.88 7 ANT3 APOA4 0.81 0.33 0.72 1.85 8 ANT3 PEDF 0.81 0.95 0.09 1.85 9 PON1 VTNC 0.59 0.81 0.45 1.84 10 CFAB HEMO 0.86 0.39 0.56 1.81 11 ANT3 HEP2 0.18 0.72 0.89 1.79 12 A1BG FETUA 0.43 0.63 0.68 1.73 13 ALBU APOA4 0.82 0.81 0.09 1.72 14 APOA1 APOM 0.39 0.57 0.75 1.71 15 CO4B ITIH2 0.63 0.77 0.30 1.71 16 A1AT FETUA 0.54 0.76 0.38 1.67 17 ALBU HEMO 0.62 0.45 0.60 1.67 18 APOB FIBA 0.67 0.51 0.49 1.67 19 HEP2 ITIH4 0.15 0.69 0.81 1.65 20 ANT3 ITIH4 -0.01 0.93 0.69 1.61 21 A1BG CFAB 0.78 0.69 0.13 1.60 22 APOA2 CFAB 0.47 0.44 0.69 1.60 23 PON1 THRB 0.70 0.37 0.53 1.59 24 ALS HEP2 0.12 0.82 0.63 1.58 25 APOA1 APOA4 0.01 0.65 0.91 1.58 26 A1AT APOA4 0.52 0.76 0.29 1.57 27 CLUS KNG1 0.28 0.70 0.59 1.57 28 ALS CLUS 0.70 0.73 0.13 1.56 29 APOC3 APOE 0.69 0.66 0.20 1.55 30 FETUA KNG1 0.61 0.18 0.76 1.55 31 APOC3 APOM 0.52 0.51 0.50 1.54 32 APOC2 APOC3 0.35 0.70 0.49 1.54 33 APOA4 CFAB 0.06 0.83 0.65 1.54 34 APOL1 HPTR 0.56 0.58 0.40 1.54 35 A1AT A2AP -0.10 0.80 0.82 1.52 36 APOA2 ITIH4 0.60 0.29 0.62 1.51 37 APOA1 ITIH4 0.56 0.36 0.57 1.50 38 THRB VTNC 0.89 0.04 0.55 1.48 39 APOA2 APOM 0.27 0.55 0.66 1.48 40 A2AP ITIH4 0.13 0.54 0.80 1.47 41 APOA4 PEDF 0.65 0.32 0.50 1.46 42 A2AP CFAB 0.56 0.39 0.51 1.46 43 A1AT ITIH4 0.04 0.56 0.85 1.45 44 A1BG CLUS 0.47 0.15 0.83 1.45 45 ALS APOA1 0.56 0.28 0.60 1.44 46 APOA1 CFAB 0.30 0.44 0.69 1.44 47 CFAB HEP2 0.96 -0.15 0.62 1.43 48 ALS APOA2 0.61 0.17 0.65 1.43 49 ALBU CFAB 0.40 0.84 0.19 1.42 50 APOE ITIH2 0.65 0.41 0.33 1.38 111

51 APOE APOM 0.50 0.43 0.44 1.38 52 A1AT A1BG 0.07 0.72 0.58 1.37 53 HEP2 THRB 0.95 -0.12 0.54 1.37 54 CLUS HEP2 0.33 0.78 0.24 1.35 55 C4BP FIBA 0.58 0.49 0.28 1.35 56 CLUS IC1 0.10 0.65 0.59 1.33 57 APOB C4BP 0.95 0.04 0.34 1.33 58 A2AP HEP2 0.45 0.12 0.75 1.32 59 A1AT ANT3 0.39 0.39 0.53 1.32 60 C1S CO3 0.73 -0.14 0.72 1.31 61 ALBU VTDB 0.57 0.80 -0.07 1.31 62 A1BG KNG1 0.44 0.08 0.77 1.30 63 APOC2 APOE 0.36 0.54 0.40 1.29 64 A1AT TTHY 0.54 -0.10 0.86 1.29 65 A2AP CLUS 0.42 0.34 0.52 1.28 66 APOA2 HEP2 0.43 0.13 0.71 1.27 67 A1AT HEP2 0.30 0.34 0.63 1.27 68 APOA1 APOC3 0.53 0.30 0.44 1.27 69 APOA2 APOA4 -0.13 0.56 0.84 1.26 70 APOA1 APOC1 0.59 0.68 0.00 1.26 71 ITIH2 PON1 -0.01 0.65 0.60 1.24 72 CFAB THRB 0.95 -0.14 0.43 1.24 73 A2AP APOA1 0.22 0.57 0.44 1.24 74 HEP2 KNG1 0.57 0.78 -0.12 1.22 75 HEP2 PON1 0.63 -0.22 0.81 1.22 76 IC1 KNG1 0.49 0.51 0.23 1.22 77 APOA4 VTDB 0.57 0.54 0.10 1.21 78 C1S ITIH2 0.11 0.76 0.32 1.20 79 A2AP ALS 0.21 0.42 0.56 1.19 80 ALS ANT3 -0.13 0.76 0.55 1.19 81 APOA1 HEP2 0.29 0.18 0.72 1.18 82 APOA2 APOC3 0.43 0.30 0.46 1.18 83 APOA4 HEMO 0.32 0.24 0.62 1.18 84 APOA2 CLUS 0.75 0.39 0.03 1.17 85 ANT3 TTHY 0.72 -0.12 0.57 1.17 86 APOA4 HEP2 0.22 0.16 0.78 1.16 87 HEP2 VTNC 0.95 -0.21 0.41 1.16 88 A1AT CLUS -0.03 0.43 0.74 1.14 89 APOE APOL1 0.31 0.09 0.73 1.14 90 A1AT VTDB 0.33 0.20 0.61 1.14 91 APOH TRFE 0.34 0.80 0.00 1.13 92 APOC3 CO3 0.61 0.63 -0.12 1.13 93 APOA1 CLUS 0.67 0.50 -0.04 1.13 94 APOC3 ITIH2 0.58 0.24 0.30 1.12 95 A2AP APOA2 0.29 0.38 0.44 1.12 96 APOC2 TTHY -0.04 0.61 0.54 1.11 97 APOC3 CLUS 0.31 0.32 0.47 1.10 98 A1AT CFAB 0.17 0.45 0.48 1.10 99 A1AT IC1 0.05 0.30 0.76 1.10 100 A2AP APOA4 -0.09 0.71 0.47 1.09 112

A B 1 1 ITIH4 0.9 HPTR 0.8 ALS 0.8 apoL1 Gel filtration 0.7 0.6 0.6 0.5 0.4 0.4 Gel filtration 0.3 0.2 0.2 0.1 0 0 Normalized abundance Normalized Normalized abundance Normalized 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 1 1 0.9 HPTR ITIH4 0.8 apoL1 0.8 ALS 0.7 0.6 0.6 0.5 0.4 Anion 0.4 Anion Exchange 0.3 Exchange 0.2 0.2 0.1

0 abundance Normalized 0 Normalized abundance Normalized 5 7 9 11 13 15 17 19 21 23 25 27 5 7 9 11 13 15 17 19 21 23 25 27 1 1 HPTR ITIH4 0.9 0.8 apoL1 0.8 ALS IEF 0.7 0.6 0.6 IEF 0.5 0.4 0.4 0.3 0.2 0.2 0.1 0 0 Normalized abundance Normalized

Normalized abundance Normalized 1 2 3 4 5 6 7 8 9 1011121314151617181920 1 2 3 4 5 6 7 8 9 1011121314151617181920 Fraction number Fraction number

Figure 3-7. Tracking co- migratory patterns of protein pairs across different separation techniques. Distribution profiles of high scoring proteins pairs across fractions from each separation technique. Panel A (left) shows the distribution traces for ITIH4 and ALS for gel filtration, anion exchange and isoelectric focusing (IEF) (top to bottom). Panel B shows the same for HPTR and apoL1, components of the biochemically confirmed HDL subspecies Trypanosome lytic factor (TLF).

113

Chapter 4. Using apolipoprotein deficient systems to study HDL composition

Introduction

In the preceding chapters of this thesis, we have discussed in detail what is known about the protein composition of HDL and described studies that we have performed to further this knowledge. The proteins found on an HDL particle likely direct the functionality of that particle, potentially making it protective against or permissive of a specific disease state such as cardiovascular disease or certain infections like those protected against by the TLF (described in Chapter 3) 1. Because of this, it is important to understand what is driving the association of specific proteins with HDL and how their association affects particle function.

The most abundant protein component of HDL is apoA-I. This protein makes up about 65 % of HDL associated protein mass when isolated by ultracentrifugation, and is often thought of as a structural scaffold upon which HDL are formed 2. Because of its abundance, many often define HDL based on the presence of apoA-I and believe that apoA-I is a strict requirement for the formation of HDL particles. The remaining protein mass of HDL is composed of apoA-II (~15%) and studies by our group and others, discussed in previous chapters, have identified as many as 50 other minor HDL associated proteins. These other minor HDL associated proteins have known functions ranging from lipid transport to regulation of the complement system and may be playing roles in currently undiscovered functions of HDL 1.

114

Currently, little is known about the basis for the association of these minor proteins with HDL i.e. whether it is mediated via interactions with other proteins on the particle or the lipid component. To begin to tease out the basis for association of minor

HDL associated proteins with lipoprotein particles we utilized mouse models with different apolipoproteins genetically ablated. Our hypothesis in this study was that many of the minor HDL proteins are associating with the particle via protein to protein interactions with one or more of the abundant proteins of HDL. To test this, we analyzed protein distribution patterns across lipoprotein subfractions from mice when the genes for three different major HDL proteins knocked out individually: apoA-I, apoA-

II and apoA-IV.

To examine this effect in the human system, we developed a collaboration with a physician who regularly sees a patient with familial apoA-I deficiency and performed a similar compositional analysis of lipoprotein subfractions. This type of genetic deficiency of apoA-I is extremely rare and has only reported in 16 families throughout the world 3. Because of the rarity of this type of mutation in humans, larger scale studies on this population are extremely difficult and we were only able to recruit one participant for this study. It is also interesting to note that although these patients have markedly reduced plasma HDL-C levels, usually in the low 20’s (mg/dL), there is no clear association with increased occurrence of cardiovascular disease 4. This is despite the fact that apoA-I containing HDL are absent in these patients. Perhaps the remaining HDL species are 1) compensating for the absence of these particles or 2) may actually be the primary mode of HDL mediated protection against cardiovascular diseases.

115

Experimental Section

Animal care. All mice strains were of C57BL/6J background. ApoA-I and apoA-II knockout mice were between 6 and 10 weeks of age with independent age matched wild type control groups. The apoA-IV knockout mice were 32 weeks old with an independent age matched control group. Blood was collected from mice by cardiac puncture using citrate as anticoagulant.

Human participants. An apoA-I deficient female participant (age 4X) was carefully paired with a control female participant (age 43). Venous blood was collected from participants after a 12 hour fast using the methods described in detail in Chapter 2.

Written consent was obtained from participants in compliance with institutional regulations.

Plasma separation by gel filtration chromatography. Plasma was obtained by centrifugation of blood at 1590xg for 15 min at room temperature, then 370 µL of plasma was applied to three superdex 200 columns (GE Healthcare) arranged in series on an Akta FPLC (GE Healthcare) as previously described in chapter 2 5. Eluate was collected as 1.5 mL fractions; these were assayed for phospholipid and cholesterol using enzymatic kits (Wako and Pointe Scientific).

116

Mass spectrometry analysis of lipid associated proteins. Lipid associated proteins were isolated from fractions using a lipid binding resin as previously characterized 5.

Proteins were trypsinized on resin and peptides collected, reduced and carboxymethylated then dried in speedvac and stored at -20 ºC until MS analysis.

Peptides were resuspended in solvent A (95% water, 5% acetonitrile, 0.1% formic acid) then 0.5 µL were injected onto a C-18 reverse phase column (GRACE; 150 × 0.500 mm) where an acetonitrile concentration gradient (5-30% in water with 0.1% formic acid) was used to elute peptides for online ESI-MS/MS by a QStarXL mass spectrometer (Applied Biosystems). For Human data, all analysis was performed as described in Chapter 2. Mouse data analysis was the same except that search criteria were performed under the taxonomy setting for mus musculus rather than homo sapiens.

Results

Apolipoprotein knockout mouse studies

Lipid distribution profiles.

ApoA-I knockout. Mouse plasma was separated by size exclusion chromatography and collected fractions were assayed for phospholipid and cholesterol content.

Consistent with measurements of total plasma PL, the profile in Fig. 4-1 shows a dramatic ~70% overall decrease in plasma PL in the apoA-I KO (area under the curve)

117 compared to WT. This decrease was primarily observed in the “HDL” fractions from 19-

25 while PL levels in the “LDL/VLDL” size range were impacted by about -50% and a minor PL peak in the small HDL size range (Fractions 27 and 28) was also decreased.

Of the HDL PL that remained, the particles appeared to be shifted toward the larger end centered on fraction 22. The cholesterol distribution was very similar to that of the PL and was decreased by 71% in the apoA-I knockout.

ApoA-II knockout. The apoA-II KO also had a massive decrease in HDL PL compared to wild type (79%, Figure 4-2). The PL that remained in these animals was associated with smaller sized particles than the remaining PL in the apoA-I knockout centered at about fraction 24. These animals also showed decreases in the large VLDL/LDL fractions and almost complete absence of PL in the small HDL fractions 27 and 28, where some had remained in the apoA-I knockout. Cholesterol was reduced by 62%, mostly in the major HDL peak across fractions 19-25.

ApoA-IV knockout. The apoA-IV KO showed little impact on the distribution profile for phospholipid or cholesterol (Figure 4-3).

An overlay of lipid traces from all groups is included as Figure 4-4.

Effects of knockout on protein distribution.

118

We next examined the effects of protein knockout on the distribution of other known lipid associated proteins. A total of 86 lipid-associating proteins were identified in all animals. By analyzing shifts in protein distribution patterns that occur in each knockout animal we can determine the importance of the knocked out protein in the formation of lipoprotein particles which contain the affected minor proteins (Figure 4-5).

ApoA-I knockout. Surprisingly, knockout of the most abundant HDL protein apoA-I had only a minor impact on the distribution of other lipid associated proteins. This is despite the dramatic reduction in overall lipid present in the HDL fractions. Only nine proteins showed unequivocal particle size shifts in the absence of apoA-I. The distribution profiles for apoA-I and a coomassie stained gel demonstrating the knockout are displayed in (Fig. 4-6). Examples of typical distribution profile shifts detected by mass spectrometry are shown in Figure 4-7. Some proteins became undetectable in the knockout such as serum amyloid A4. For others one peak was absent as in paraoxonase 1 (PON1) or whole peaks were shifted as in phosphatidylinositol-glycan- specific phospholipase-D (Figure 4-7).

ApoA-II knockout. Knockout of apoA-II had a dramatic effect on other lipid associated proteins resulting in shifts for 46 proteins. Examples of shifted protein distributions are shown in Figure 4-8. Reduced spectral counts for apoA-I were detected across the major peak, in fractions containing large HDL. This is consistent with our phospholipid and cholesterol distribution data which showed the greater reduction in the large HDL fractions. Other proteins shifted in ways similar to those observed in the apoA-I

119 knockout having complete shifts in distribution or no longer associating with lipid and becoming undetectable such as Complement C2. The fact that the number of affected proteins in the apoA-II knockout is five times higher than the number of proteins affected by knockout of apoA-I was highly surprising (Fig. 4-10 and Table 4-1). Given the great body of literature supporting apoA-I as an essential component of HDL structure and the relatively minor role of apoA-II, we might have expected the opposite to be true. It may be that although apoA-II is not necessarily required for the formation and stability of

HDL, it is playing an extremely important role in the association of many of these minor

HDL proteins with the already formed particle.

ApoA-IV knockout. Knockout of apoA-IV influenced the shift of 9 proteins having only

1 in common with the apoA-I knockout, this was apoE. Examples of protein shifts are shown in Figure 4-9. In this model, the effects on protein distribution seemed minor in comparison to those seen in apoA-I and apoA-II knockout models. This is consistent with the very small impact observed in the lipid profile distributions and suggests a minor role for apoA-IV in the association of minor HDL proteins with the particle.

ApoA-I deficient human study

To take the next step in this study, we performed a similar analysis in a human participant. Because of obvious limitations in this type of genetic analysis in humans we

120 were lucky to meet a contact with regularly saw a patient with genetic deficiency of apoA-I.

Phospholipid distributions for apoA-I deficient and control participants are shown in Figure 4-11. As expected, the apoA-I deficient subject exhibited a large reduction in phospholipid associated with HDL size fractions (fractions 20 – 30). Comparing this to the apoA-I KO mice in Figure 4-1 indicates that loss of apoA-I had a greater impact on the LDL/VLDL fractions in the mouse vs. the human. Also, the mice tended to have greater residual phospholipid remaining in the HDL fractions compared to the human which exhibited only minimal detectable PL in these fractions.

Mass spectrometry was used to determine the relative distribution patterns for proteins across collected fractions for the two participants. A total of 41 known HDL associated proteins were identified. Four of these were detectable only in the control participant: apoA-I, apoL-I, haptoglobin related protein, and serum amyloid A4 (Figure

4-12). Of the 41 detected proteins 18 were determined to shift in distribution pattern as a result of apoA-I deficiency. Representative examples of distribution profiles for shifted proteins are shown in Figures 4-13 a – c and an example of an unaffected protein distribution profile is shown in Figure 4-13 d. Table 4-2 is a list of all identified proteins and their shift status.

Discussion

The mouse model represents an invaluable system for the study of genetic effects on lipoprotein metabolism. Their plasma contains lipoproteins that are roughly similar to

121 those in the human as far as protein and lipid composition and density distribution 6.

That is they have VLDL density particles that are rich in triglyceride and LDL which are rich in cholesterol and both contain a protein of similar size and immunoreactivity to human apoB. They also have lipoprotein particles in the density range of human HDL which contain apoA-I, apoA-II, apoE and other known human HDL associated proteins.

One key difference between the mouse and human system is the distribution of these lipoprotein components. In the human, low density lipoproteins represent the majority of circulating lipoproteins however the mouse has an HDL centric system, with LDL only accounting for about 20% of lipoproteins 6. These high levels of HDL and low LDL may partially account for the lack of naturally occurring cardiovascular disease in these animals.

Many in the field think of apoA-I as an absolute requirement for the formation and stability of HDL in the circulation. This study demonstrates that even in the absence of apoA-I, and with the resultant ~70% decrease in plasma phospholipid content in the mouse, the majority of the minor HDL associated proteins still associate with phospholipid and migrate to fractions representing HDL sized particles. Just as interesting is the fact that knockout of apoA-II has a similar effect on HDL phospholipid levels but has a much greater impact on the migratory patterns of minor HDL proteins suggesting that apoA-II plays an important role in a) stabilizing all HDL, even those with apoA-I, b) may form a scaffold for some of the more minor proteins and its disruption dramatically affects them, or c) loss of apoA-II affects lipoprotein metabolism in such a way that lipoprotein profiles are altered. There is some evidence for this last option.

Previous studies indicate that apoA-II may play in important role in the maintenance of

122 plasma HDL levels by inhibiting the enzyme hepatic lipase, which hydrolyzes HDL triglycerides and results in smaller HDL 7. Therefore one potential mechanism for disruption of many of the HDL protein associations observed here could be that the absence of apoA-II is resulting in increased hepatic lipase activity. This increased activity may be changing the particle size and structure in such a way that these proteins no longer have an affinity to associate. Knockout of apoA-IV had little effect on less than 10% of identified proteins suggesting only a minor role structurally or metabolically in the association of protein components with HDL.

The data presented here provide an interesting step forward in our understanding of protein to protein interactions on lipoprotein particles. Our hypothesis at the beginning of this study was that many of the minor HDL associated proteins associate with the particle by interacting with some of the most abundant proteins which are likely to be found on the majority of HDL particles. Our results indicate that for many proteins this may be the case and that of the proteins studied here apoA-II may be a major factor in the association of these other proteins with HDL. What was most surprising to us was the lack of a major effect on protein distribution patterns when apoA-I was absent.

This lack of effect was quite pronounced in the mouse with only about 10% of identified proteins displaying distribution shifts. When translated to the human system the absence of apoA-I affected about 44% of the other proteins, this is greater than in the mouse but still very low if you accept the premise that apoA-I is required for the formation of HDL particles. In fact, if this were the case we could expect to collect no lipid or lipid associated protein in the size range of HDL (fractions 20 – 30). Because this is not the case we conclude from this data that apoA-I is not strictly required for the

123 formation of all lipidated particles in the HDL size range. Comparing this study in humans versus the mouse study we find that of the 9 proteins identified to shift in the mouse apoA-I knockout model, 8 were also identified in the human and of these 4 were also determined to shift in the human while 4 of them did not. For those 4 proteins that shifted commonly between the mouse and human study (apoC-I, apoC-III, albumin, and

PON1), this data provides convincing evidence for an important role of apoA-I in the formation of particles which contain these proteins.

We can also compare the results of this study to those from our correlation analysis in Chapter 3 where we identified pairs of proteins likely to reside together on a single HDL particle. Again, we are comparing the mouse against a human system so cannot expect total agreement but here we do find support for several of our proposed protein interactions. In fact more than half (5 out of 9) of the proteins found to shift in the apoA-I knockout mouse were consistent with protein pairings with apoA-I which scored in the top 25% of the correlation analysis results (Table 4-3). Additionally, about

24% of proteins shifted in the apoA-II knockout model were proposed to associate with apoA-II in the top 25% of our correlation analysis in humans. However, none of the proteins shifted in the apoA-IV knockout model were supported. This suggests a good degree of similarity between protein associations with HDL in the mouse and human systems and provides experimental evidence in support of our correlation analysis.

An interesting observation from the apoA-I deficient human study that was not seen in the mouse system was the lack of proteins apoL-I and haptoglobin related protein. These proteins are the components of an HDL subspecies known as TLF, described in detail in Chapter 3. Our data suggest the absence of this particle in apoA-I

124 deficient persons, consistent with the fact that apoA-I is also an important component of this this subspecies. The other undetected protein in the apoA-I deficient participant was serum amyloid A4 (SAA4). This protein is constitutively expressed by the liver and secreted into the circulation where it is known to associate with lipoproteins but its physiological functions are largely unknown.

This human data presented in this study has limitations imposed by very small sample size and the high degree of variability inherent in human subjects. These obstacles are nearly impossible to overcome due to the highly infrequent occurrence of the particular types of specific genetic mutations required for these studies. We did our best to select an appropriate control participant who was matched for age and sex and other variables which may affect lipid profile.

From these studies it is not possible to tell whether the mechanism of particle disruption is a direct structural effect or indirect result of changes in metabolism that occurred due to the chronic absence of a given apolipoprotein. Two potential mechanisms of a direct disruption are that the knockout resulted in a disruption in protein to protein interactions on the particle or that absence of the given protein resulted in changes in surface curvature of the affected particles potentially disrupting normal protein to lipid interactions. Another possibility could be that whole body knockout of these important apolipoproteins is causing changes in metabolism that are indirectly affecting particle structure. Overall these data suggest that apoA-I, while associated with the majority of HDL phospholipid, may not be required for the formation of many minor HDL subspecies. ApoA-II on the other hand, may be involved in the formation of many of these, perhaps acting as a docking site or adaptor molecule. This

125 study will provide a basis for the further study of HDL subspecies and the was that proteins interact on their surface and set more directed studies targeting specific protein to protein interactions in in vitro systems that will provide direct evidence of interaction and possible functional consequences of those interactions.

126

Reference List

1 Gordon SM, Hofmann S, Askew DS, Davidson WS. High density lipoprotein: it's

not just about lipid transport anymore. Trends Endocrinol Metab 2011;22(1):9-15.

2 Davidson WS, Thompson TB. The structure of apolipoprotein A-I in high density

lipoproteins. J Biol Chem 2007;282(31):22249-53.

3 Al-Sarraf A, Al-Ghofaili K, Sullivan DR, Wasan KM, Hegele R, Frohlich J. Complete

Apo AI deficiency in an Iraqi Mandaean family: case studies and review of the

literature. J Clin Lipidol 2010;4(5):420-6.

4 Santos RD, Schaefer EJ, Asztalos BF, Polisecki E, Wang J, Hegele RA, Martinez

LR, Miname MH, Rochitte CE, Da Luz PL, Maranhao RC. Characterization of high

density lipoprotein particles in familial apolipoprotein A-I deficiency. J Lipid Res

2008;49(2):349-57.

5 Gordon SM, Deng J, Lu LJ, Davidson WS. Proteomic characterization of human

plasma high density lipoprotein fractionated by gel filtration chromatography. J

Proteome Res 2010;9:5239-49.

6 Camus MC, Chapman MJ, Forgez P, Laplaud PM. Distribution and

characterization of the serum lipoproteins and apoproteins in the mouse, Mus

musculus. J Lipid Res 1983;24(9):1210-28.

127

7 Weng W, Brandenburg NA, Zhong S, Halkias J, Wu L, Jiang XC, Tall A, Breslow

JL. ApoA-II maintains HDL levels in part by inhibition of hepatic lipase. Studies In

apoA-II and hepatic lipase double knockout mice. J Lipid Res 1999;40(6):1064-70.

128

0.1 0.05 WT (n=6) WT

0.08 A-I KO (n=3) 0.04 A-I KO

0.06 0.03

0.04 0.02

Figure 4-1. Lipid distribution profiles for apoA-I knockout mice. Distribution of phospholipid (top) and cholesterol (bottom) across plasma fractions collected by gel filtration chromatography.

129

0.1 0.05 WT (n=6) WT

0.08 A-II KO (n=4) 0.04 A-II KO

0.06 0.03

0.04 0.02

0.02 0.01 Phospholipid (mg/mL) Phospholipid Cholesterol (mg/mL) Cholesterol 0 0 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Fraction Number Fraction Number Figure 4-2. Lipid distribution profiles for apoA-II knockout mice. Distribution of phospholipid (top) and cholesterol (bottom) across plasma fractions collected by gel filtration chromatography.

130

0.1 0.05 WT (n=6) WT

0.08 A-IV KO (n=3) 0.04 A-IV KO

0.06 0.03

0.04 0.02

0.02 0.01 Phospholipid (mg/mL) Phospholipid Cholesterol (mg/mL) Cholesterol 0 0 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Fraction Number Fraction Number Figure 4-3. Lipid distribution profiles for apoA-IV knockout mice. Distribution of phospholipid (top) and cholesterol (bottom) across plasma fractions collected by gel filtration chromatography.

131

Figure 4-4. Lipid distribution profiles for apolipoprotein knockout mice. Distribution of phospholipid (top) and cholesterol (bottom) across plasma fractions collected by gel filtration chromatography.

132

Figure 4-5. Strategy for identifying proteins affected in knockout models.

133

Figure 4-6. Characterization of apoA-I knockout mice. Distribution of phospholipid associated apoA-I protein across gel filtration fractions detected by mass spectrometry (left). Right, coomassie stained SDS gel of pooled plasma mouse from apoA-I knockout (KO) and wildtype control (Wt) mice.

134

Figure 4-7. Distribution profiles of some proteins shift as a result of apoA-I knockout.

135

apoA-I apoE 25 5 WT 4.5 WT 20 A-II KO 4 apoA-II KO 3.5 15 3 2.5 10 2 1.5 5 1

Spectral counts Spectral counts Spectral 0.5 0 0 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Fraction number Fraction number

Complement C2 Complement C4-B 1.4 12 A-II KO A-II KO 1.2 10 Wildtype Wildtype 1 8 0.8 6 0.6 4 0.4

Spectral counts Spectral 0.2 counts Spectral 2 0 0 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Fraction number Fraction number

Figure 4-8. Distribution profiles of some proteins shift as a result of apoA-II knockout.

136

apoA-I apoE 25 6 WT WT 5 20 apoA-IV KO apoA-IV KO 4 15 3 10 2 5 1 Spectral counts Spectral counts Spectral 0 0 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Fraction number Fraction number

Coagulation Factor XIII B Vitronectin 1.4 7 WT WT 1.2 6 apoA-IV KO apoA-IV KO 1 5 0.8 4 0.6 3 0.4 2 0.2 1 Spectral counts Spectral counts Spectral 0 0 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Fraction number Fraction number

Figure 4-9. Distribution profiles of some proteins shift as a result of apoA-IV knockout.

137

Figure 4-10. Number of protein shifts detected in each knockout model.

138

Table 4-1. List of proteins with shifted distribution profiles in each knockout model.

139

Figure 4-11. Phospholipid distribution profiles for control and apoA-I deficient human plasma fractions collected by gel filtration chromatography.

140

Figure 4-12. Venn diagram displaying total identified protein numbers between control and A-I deficient human plasma fractions.

141

Figure 4-13. Comparison of protein distribution profiles. Representative examples of proteins which shifted in apoA-I deficient subject compared to control (a-c) and those that did not shift (d).

142

Table 4-2. List of all identified proteins and whether or not they were shifted in apoA-I deficiency.

143

Table 4-3. Correlation analysis results supported by mouse knockout studies. Correlation Score Rank Protein A Protein B GF AE IEF Combined 1 APOA1 APOA2 0.89 0.80 0.86 2.55 48 ALS APOA2 0.61 0.17 0.65 1.43 68 APOA1 APOC3 0.53 0.30 0.44 1.27 69 APOA2 APOA4 -0.13 0.56 0.84 1.26 70 APOA1 APOC1 0.59 0.68 0.00 1.26 82 APOA2 APOC3 0.43 0.30 0.46 1.18 84 APOA2 CLUS 0.75 0.39 0.03 1.17 95 A2AP APOA2 0.29 0.38 0.44 1.12 106 APOA1 APOE 0.17 0.15 0.74 1.06 132 APOA1 HEMO 0.28 0.02 0.66 0.96 142 APOA2 APOC1 0.52 0.56 -0.15 0.93 145 A1AT APOA2 0.10 0.46 0.36 0.92 164 APOA2 TTHY -0.11 0.40 0.54 0.83 180 APOA2 PON1 0.69 -0.32 0.42 0.79 181 APOA1 PON1 0.60 -0.30 0.48 0.79 189 APOA2 APOE 0.09 0.12 0.56 0.77

144

Chapter 5. Functional analysis of HDL subspecies

Introduction

In this chapter we move from studies of HDL composition to studies of HDL function.

Previous functional studies on pools of total HDL isolated from humans have identified a variety of functional attributes for these dynamic lipoproteins many of which are thought to provide protection against cardiovascular disease. These functions include HDL’s role in reverse cholesterol transport (i.e. CH efflux), anti-inflammation and anti-oxidation.

The known mechanistic details behind HDL’s ability to perform these functions are discussed in Chapter 1 so we will not mention them in great detail again here 1.

Aim 1 of this thesis hypothesized that the total pool of HDL in an individual was composed of numerous distinct subspecies with varying protein compositions. Our experiments addressing Aim 1 in Chapters 2-4 have supported this idea.. The hypothesis driving Aim 2 is that the compositional diversity of HDL may result in functional heterogeneity among subspecies. Indeed some examples of protein function on HDL are known, such as the anti-oxidative activity of paraoxonase 1 (PON1) 2 and the lipid transfer activities of CETP and LCAT 3. However, we do not know the extent to which these proteins contribute to a given activity on HDL or if other proteins also contribute. Additionally, we do not know which subspecies of HDL contain these proteins and perform these functions.

145

To assess functional diversity across HDL subspecies we performed a panel of functional assays on lipoprotein fractions isolated using the gel filtration chromatography methods developed in Chapter 2. We chose assays for the most well understood of

HDL functions that are thought to contribute to protection against cardiovascular disease: cholesterol efflux, inhibition of LDL oxidation and anti-inflammation. Our previous experiments have provided a large amount of compositional data about the way proteins distribute across HDL fractions collected by the gel filtration technique.

Here we aim to collect functional distribution data across fractions from these same subjects with the goal of correlating this data with the compositional data to identify proteins that may be involved in each function.

Methods

Gel filtration fractionation of human plasma. This was performed exactly as described in Chapter 2 however buffers were prepared without sodium azide and EDTA to minimize harmful effects on cell culture systems or experiments which depend on oxidation chemistry. Additionally, the same three normolipidemic male participants whose lipoprotein proteome was characterized in Chapter 2 were used for these functional assays to allow for correlation between protein composition and functional data.

Cholesterol efflux. J774 murine macrophage cells were seeded into 24 well plates and grown until 80% confluent then incubated in 1 mL of labeling media overnight

146

(RPMI+0.2% BSA+0.3 mM cyclic AMP+radiolabeled cholesterol). During labeling, expression of HDL efflux transporter ABCA1 is upregulated by incubation with cyclic

AMP. Cells were washed three times with RPMI+0.2% BSA then lipid containing fractions were added at constant volume (50 uL/well) to a total volume of 1 mL and incubated for 6 hours at 37ºC. Each fraction from all subjects were assayed in triplicate.

One plate, designated as baseline (T0), was used to determine initial radiocholesterol loading prior to fraction treatment. These wells were washed with PBS three times and cells were lysed with isopropanol and radioactivity measured. To measure efflux from cells to the media, media was collected from the sample plates (Ts), filtered to remove any cells and radioactivity was counted in a scintillation counter. Cholesterol efflux was calculated as: %Efflux = (Ts(avg)/T0(avg))*100.

Inhibition of LDL oxidation. LDL was isolated from 3 units of human plasma obtained from Hoxworth blood bank (Cincinnati, OH). LDL (20 ug cholesterol) was incubated in

PBS with AAPH (2,2'-azobis-2-methyl-propanimidamide, dihydrochloride) an azo compound which generates reactive oxygen species at a constant rate in water and collected gel filtration fractions or buffer alone. This mixture was monitored at 234 nm in a plate reader for 6 hours while maintained at 37ºC with intermittent shaking. Increasing absorbance at 234 nm over time indicates the accumulation of conjugated diene lipid oxidation products in the samples. The rate of oxidation was determined by measuring the slope of the absorbance curve. The percent change in rate of oxidation (% change

A234 slope) was calculated as: (% change A234 slope) = 100 – ((slope treated

147 fraction/slope of buffer control)*100). Therefore, a negative value indicates a reduction in the rate of oxidation in the presence of a given fraction.

Anti-inflammation. Assays of inflammation were done in collaboration with Dr. Kasey

Vickers (NIH; Bethesda, MD). Collected fractions were incubated with Human coronary artery endothelial cells plated at 1x 10^5 cells/mL for 24 hours. Cells were treated with

900 µL EBM-2 serum free media + 200 µL of concentrated fraction for 24 hours. Cells were washed two times with PBS. Then total RNA was isolated using Qiagen miRNA

Easy Mini Kits RT and real-time PCR was used to measure RNA for ICAM1, VCAM1, and PRMD1 and normalized to PPIA. The profile of ICAM1 and VCAM1 mRNA fold changes were combined to represent an “Activation” profile representative of the endothelial cells inflammatory state based on these markers. This assay was only performed on one of the three subjects.

Functional correlation analysis. For each of the three functions, we used Pearson

Correlation to calculate the similarity between activity profile and each protein’s

th abundance profile (Eq.2). Fi is the function activity value for the given function in the i

th fraction, and Xi is the protein X’s abundance value in the i fraction.

()()FFXX i ii rFX,  (Eq.2) ()FF22 ( X X ) iiii

Results

148

Cholesterol efflux

We measured the ability of individual fractions from 3 healthy donors to efflux cholesterol from ABCA1 activated macrophages. The averaged data is shown in Figure

5-1. The distribution of efflux activity appears as two distinct peaks the first across larger size fractions 15 through 18, which we have previously characterized as containing VLDL and LDL. This peak makes up approximately 50% of the total area under the trace, equal to the second peak which is spread across fractions 20 through

28 encompassing the size range of HDL particles. It is interesting to note that this distribution pattern almost perfectly mirrors the distribution of phospholipid across fractions collected from a healthy subject (see Chapter 2). This suggests that perhaps efflux activity in this assay is driven in a large part by phospholipid content, which is consistent with previous observations 4.

Inhibition of LDL oxidation

Next, we measured the ability of HDL to inhibit AAPH induced lipid oxidation of LDL.

This data is shown in Figure 5-2; here we can see that the distribution pattern of activity for this function is quite different from that of cholesterol efflux. Anti-oxidative capacity of HDL as measured by this assay is most potent in those fractions containing the smallest HDL (fractions 27–29). The fraction with the strongest anti-oxidative function is fraction 28 which induced a 70% decrease in the rate of LDL oxidation.

Anti-inflammatory properties

149

The ability of HDL subfractions to inhibit inflammation was measured by monitoring changes in the expression of cell adhesion molecules by endothelial cells. Data was expressed as fold change in activation (expression) where increased activation is pro- inflammatory. We found that while some fractions, 26–27, reduced activation by over

50%, fractions 18 and 21–23 actually promoted activation, some by over 2 fold in the large HDL size range (Figure 5-3). While it is known that total HDL can have potent anti-inflammatory properties, there do exist so called pro-inflammatory HDL 5 that are less well-studied and may represent what we have isolated in these large HDL fractions.

Correlation analysis

To determine if specific protein components of the HDL are responsible for the functional properties studied in the previous sections of this chapter we used a correlation analysis. This analysis compared the functional distribution profiles created in this study (Figure 5-4) to the proteomic distribution profiles obtained from the same subjects and reported in Chapter 2. Table 5-1 displays the averaged protein distribution profiles for these subjects as a heat map and also displays the correlation coefficient values for the correlation of each given protein with each function, significant correlation values are highlighted in red.

Cholesterol efflux activity correlated significantly with four proteins: apoB, fibrinogen alpha chain, and complement C1q subunits B and C which were found primarily in larger LDL containing fractions 15–18 (Table 5-1). No significant correlations between cholesterol efflux and any protein were identified across the HDL size range fractions. The lack of correlation with protein content supports a lipid based

150 driving force for this function in these fractions, as suggested by the high degree of similarity between phospholipid distribution patterns and cholesterol efflux activity

(Figure 5-1).

The ability to inhibit oxidation of LDL was significantly correlated with 13 proteins.

Eight of these were found in the smaller HDL fractions 24–30 (Table 5-1) which accounted for over 60% of total activity in this assay. We found that some of these proteins have already been reported to possess antioxidant activity. For example, apoA-IV was the strongest correlating anti-oxidant protein in this study and has previously been shown to inhibit LDL oxidation in an assay system very similar to that described here 6. The normalized distribution profile for apoA-IV is overlayed onto

Figure 5-2. Other proteins whose distribution profiles correlated with anti-oxidative capacity in this assay were collected in larger size fractions and include inter-alpha trypsin inhibitor heavy chains 1 and 2 and apoE. ApoE is of high interest as it is already known to possess antioxidant activity that is dependent on which of three genetic variants are present E2>E3>E4 7.

Anti-inflammatory activity as measured here was correlated with the distribution of 15 proteins. Eleven of these proteins showed a positive correlation while only four were associated with decreased activation. Proteins that increased activation were found in fractions 21–23 and included apoM, ITIH3, apoC-I and C-III as well as several complement factors (Table 5-1). Those that decreased activation were found in fractions 26 and 27 and included alpha-2-HS-glycoprotein, hemopexin, alpha-1- antitrypsin and vitronectin. The strongest correlation of these was with alpha-2-HS- glycoprotein (Fetuin A), which correlated significantly with reduction in both ICAM and

151

VCAM expression. The distribution pattern of this protein is overlayed onto Figure 5-3 in red.

Discussion of functional studies

Here we have set out to assign functional roles to gel filtration isolated subfractions of

HDL. We have focused on the best established functions that have been previously measured in preparations of total HDL, most often isolated by ultracentrifugation. We acknowledge that additional functions for HDL exist that were not examined here.

These additional functions may also have important roles in protection against cardiovascular and other disease states and are still in need of further exploration. We have utilized a gel filtration chromatography technique that we developed previously to study protein distribution across size fractionated HDL. This separation technique has important advantages over the commonly used ultracentrifugation methods. First, our gel filtration approach has been optimized to collect total HDL across a large number of fractions providing a higher resolution separation of subpopulations of HDL. This is important because the focus of our study is on the functionality of HDL subspecies, not total HDL. Second, is the fact that this chromatography based technique allows for separation under physical and chemical conditions that are much more physiologically relevant than those of ultracentrifugation. As mentioned in previous chapters, ultracentrifugation involves high g forces and high salt concentrations that have been demonstrated to negatively affect HDL structure and functionality. These facts make ultracentrifugation unsuitable for the present studies. An additional goal of this study

152 was to correlate the measured functions with protein composition previously measured in the same participants studied here (Chapter 2).

The results of our functional studies demonstrate a varied distribution of functional activity across collected fractions. We have previously demonstrated the compositional diversity across lipoproteins fractionated in this manner. So the results of this study tell us that not all HDL are functionally equal. Some of the measured functions are distributed broadly across fractions with distinct peaks of activity while others are much more isolated (Figure 5-4). Some functions of HDL can even work in two directions, where different HDL species may have opposing effects i.e. pro vs anti- inflammatory HDL (Figure 5-3). The balance between these subpopulations may be an important factor in assessing cardiovascular risk associated with HDL.

Of methodological note is the fact that we chose to use equal fraction volumes in our functional assays rather than doing a pre-assay normalization to equalize based on one of the chemical components of HDL as some have done previously. It is difficult to obtain agreement among colleagues in the field on which component of HDL would be best to normalize to, if any. Some would say cholesterol, others total protein, while we would say total phospholipid content. Any one could introduce bias in the assay if the chosen component had influence on a given assay, and this influence is what we are trying to examine with these experiment in the first place. Additionally, technical issues would arise due to the wide diversity of HDL composition. One fraction might have 20 times more lipid or protein than another making it difficult to equalize volumes without difficult concentration steps prior to each assay. In the end, we found it most logical to compare the fractions on an equal volume basis. The fact that we demonstrated that

153 some fractions perform certain functions better than others supports our approach. For example, if all assays that we performed had corresponded with those fractions that contain the most PL or apoA-I, then it could be argued that all HDL is the same and, as expected, the fractions that had the most HDL present did the best in the assays.

However, our results did not show this. For example, small HDL particles were clearly more effective at preventing LDL oxidation, despite the fact that the total PL in these fractions represents only a minor component of the larger HDL peak in the gel filtration analysis. This suggests two possibilities, i) there are more particles in these fractions that are antioxidative (per unit volume), or ii) there are few particles in these fractions that have a remarkable antioxidative capacity, i.e. they have a high specific activity. In either case, normalization of the particles prior to the assay would likely show similar results provided a satisfactory component can be agreed upon for the normalization.

Some of the functional data reported here are supported by previous studies.

For example the data from our anti-oxidation assay correlates well with previous anti- oxidation studies on HDL isolated by ultracentrifugation. If we assume that with decreasing size we have a concurrent increase in density in our HDL fractions due to an increasing protein to lipid ratio inherent in smaller HDL particles, then our findings of overall increasing capacity from fractions 20–29 match those found by Kontush et al where they report increasing anti-oxidative capacity from five density fractions, HDL 2a to 3c, using a similar assay 8.

Overall, this study provides clear evidence for the functional heterogeneity of

HDL subspecies. This is further supported with correlative data suggesting possible protein or lipid bases for this heterogeneity. This study opens the door for further

154 functional analyses and experiments using reconstituted HDL particles of any desired composition to directly test the influence of specific proteins or combinations of proteins on functionality. HDL represents an extremely diverse population of particles compositionally and functionally, likely even much more diverse than we currently know.

Further exploration of the functionality of HDL subspecies and which components drive these functions will be invaluable in the successful development of HDL as a therapeutic target and biomarker for cardiovascular disease.

155

Reference List

1 Gordon SM, Hofmann S, Askew DS, Davidson WS. High density lipoprotein: it's not

just about lipid transport anymore. Trends Endocrinol Metab 2011;22(1):9-15.

2 James RW, Deakin SP. The contribution of high density lipoprotein apolipoproteins

and derivatives to serum paraoxonase-1 activity and function. Adv Exp Med Biol

2010;660:173-81.

3 Cucuianu M, Coca M, Hancu N. Reverse cholesterol transport and atherosclerosis.

A mini review. Rom J Intern Med 2007;45(1):17-27.

4 Rothblat GH, de la Llera-Moya M, Atger V, Kellner-Weibel G, Williams DL, Phillips

MC. Cell cholesterol efflux. Integration of old and new observations provides new

insights. J Lipid Res 1999;40(5):781-96.

5 Ansell BJ, Fonarow GC, Fogelman AM. The paradox of dysfunctional high-density

lipoprotein. Curr Opin Lipidol 2007;18(4):427-34.

6 Wong WMR, Gerry AB, Putt W, Roberts JL, Weinberg RB, Humphries SE, Leake

DS, Talmud PJ. Common variants of apolipoprotein A-IV differ in their ability to

inhibit low density lipoprotein oxidation. Atherosclerosis 2007;192(2):266-74.

7 Miyata M, Smith JD. Apolipoprotein E allele-specific antioxidant activity and effects

on cytotoxicity by oxidative insults and beta-amyloid peptides. Nat Genet

1996;14(1):55-61.

8 Kontush A, Chantepie S, Chapman MJ. Small, dense HDL particles exert potent

protection of atherogenic LDL against oxidative stress. Arterioscler Thromb Vasc

Biol 2003;23(10):1881-8.

156

20 0.06 18 16 0.05 14 0.04 12 10 0.03 8 Efflux % Efflux 6 0.02 4 0.01

2 (ug/uL) Phospholipid 0 0 131415161718192021222324252627282930 Fraction

Figure 5-1. Cholesterol efflux capacity across gel filtration fractions. Cholesterol efflux from cAMP activated J774 murine macrophage cells to equal volumes of gel filtration fractionated plasma was measured (solid blue trace). This activity correlated most strongly with phospholipid content (dashed red trace).

157

40 1.2

20 1

0 0.8 IV -20 0.6 -

-40 0.4 apoA oxidation -60 0.2 % change in rate of LDL LDL of rate in change % -80 0 of distribution Normalized 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

Fraction

Figure 5-2. Anti-oxidative capacity across gel filtration fractions. We measured the ability of equal volumes of lipid containing gel filtration separated plasma fractions to inhibit AAPH induced oxidation of LDL lipids (solid blue bars). This function correlated most strongly with the distribution of apolipoprotein A- IV (dashed red trace).

158

2.5 1.2

2 1 0.8

1.5 0.6 1 Change

0.4 A Fetuin

0.5 0.2

0 0 of distribution Normalized Activation (ICAM+VCAM) Fold (ICAM+VCAM) Activation 13 15 17 19 21 23 25 27 29 Fraction

Figure 5-3. Anti-inflammatory capacity across gel filtration fractions. We measured the effect of fractions on the expression of endothelial cell adhesion molecules involved in inflammation. The blue trace shows the fold change in Activation, a measure derived from the combined data on both ICAM and VCAM expression, determined by PCR after 24 hour incubation with fractions. This activity correlated most strongly with the distribution of the protein alpha-2-HS-glycoprotein (Fetuin A, dashed red trace).

159

45 LDL Oxidation 40 CH Efflux

35 Anti-inflammation 30 25 20 15 % Activity % 10 5 0 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Fraction Number Figure 5-4. Summary of functional distribution profiles.

160

Vitamin D-binding protein Transthyretin Pigment epithelium-derived factor Antithrombin-III Kallistatin Apolipoprotein A-IV Alpha-1-antitrypsin Vitronectin Kininogen-1 Heparin cofactor 2 Hemopexin Alpha-2-HS-glycoprotein Beta-2-glycoprotein 1 (apo H) Tetranectin Prothrombin Complement factor B Alpha-2-antiplasmin Alpha-1-antichymotrypsin Lumican precursor Serum paraoxonase/arylesterase 1 Inter-alpha-trypsin inhibitor heavy chain H4 Insulin-like growth factor-binding protein Complement C2 Clusterin precursor (apo J) Apolipoprotein A-II Apolipoprotein A-I Complement C3 Complement subcomponentC1s Complement precursorC5 Complement C4-B Apolipoprotein C-III Apolipoprotein C-I Ig lambda chain C regions Inter-alpha-trypsin inhibitor heavy chain H3 Plasma protease inhibitorC1 Apolipoprotein M Apolipoprotein E Inter-alpha-trypsin inhibitor heavy chain H2 Inter-alpha-trypsin inhibitor heavy protein chain H1 Haptoglobin-related Apolipoprotein-L1 Ficolin-3 Alpha-2-macroglobulin Complement subcomponentC1q subunit C Fibrinogen alpha chain Complement subcomponentC1q subunit B Apolipoprotein B-100 Protein name Table 5-1. Heat map of protein distribution and functional correlation analysis

0.03 0.02 13 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.08 0.14 0.13 0.16 0.25 0.07 0.21 0.08 0.14 0.15 0.16 0.09 0.06 0.12 0.18 0.22 0.08 0.01 0.18 0.09 0.17 0.15 0.01 0.2 14 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.12 0.25 0.14 0.13 0.16 0.13 0.07 0.11 0.12 0.05 0.11 0.09 0.03 0.12 0.12 0.23 0.03 0.04 0.33 0.27 0.82 0.09 0.17 0.11 0.67 0.27 0.53 0.6 0.5 15 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.12 0.14 0.12 0.06 0.16 0.07 0.09 0.08 0.12 0.01 0.36 0.02 0.05 0.17 0.56 0.3 16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0.08 0.07 0.38 0.21 0.21 0.32 0.08 0.21 0.17 0.25 0.19 0.09 0.13 0.11 0.27 0.24 0.34 0.01 0.09 0.97 0.33 0.32 0.2 0.5 17 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0.31 0.06 0.24 0.03 0.27 0.59 0.74 0.17 0.14 18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0.33 0.54 0.06 0.28 0.01 0.08 0.09 0.27 0.26 0.27 0.33 0.28 0.38 0.22 0.2 19 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0.67 0.13 0.21 0.11 0.08 0.03 0.09 0.03 0.46 0.35 0.54 0.15 0.64 0.17 0.91 0.67 0.33 0.12 0.1 0.4 20 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0.05 0.33 0.06 0.86 0.13 0.38 0.65 0.74 0.55 0.89 0.88 0.33 0.01 0.1 0.8 21 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 Fraction number 0.17 0.06 0.09 0.38 0.35 0.71 0.98 0.92 0.64 0.44 0.5 0.6 0.5 0.5 0.6 22 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0.04 0.19 0.33 0.06 0.14 0.16 0.16 0.07 0.14 0.27 0.25 0.42 0.38 0.82 0.99 0.36 0.18 0.67 0.64 0.18 0.19 0.15 0.13 0.11 0.1 0.2 0.5 0.6 23 0 0 0 0 0 0 0 0 1 1 0 0 1 0 0 0 0 0 0 0.12 0.14 0.21 0.33 0.31 0.07 0.26 0.36 0.08 0.21 0.25 0.19 0.45 0.65 0.02 0.67 0.45 0.09 0.17 0.13 0.15 0.05 0.2 0.6 24 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0.09 0.75 0.43 0.89 0.84 0.38 0.29 0.75 0.31 0.26 0.77 0.82 0.71 0.16 0.04 0.33 0.36 0.21 0.17 0.14 0.4 0.2 25 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0.14 0.25 0.67 0.85 0.89 0.73 0.75 0.35 0.59 0.56 0.03 0.09 0.03 0.1 0.5 0.4 26 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.04 0.25 0.71 0.75 0.91 0.63 0.43 0.58 0.48 0.85 0.43 0.35 0.38 0.25 0.12 0.12 0.41 0.03 27 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.42 0.25 0.76 0.86 0.67 0.31 0.07 0.16 0.38 0.07 0.03 0.03 0.06 0.16 0.4 0.3 28 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.38 0.64 0.06 0.11 0.28 0.15 0.03 0.18 0.13 0.11 0.08 0.18 0.26 29 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 VCAM 0.143 0.246 0.138 0.375 0.432 0.280 0.386 0.520 0.000 0.480 0.502 0.554 0.451 0.271 0.402 0.401 0.135 0.171 0.046 0.012 0.269 0.269 0.050 0.161 0.457 0.551 0.746 0.737 0.652 0.652 0.601 0.603 0.700 0.580 0.483 0.654 0.219 0.326 0.250 0.078 0.355 0.142 0.108 0.112 0.003 0.027 0.144 0.229 0.293 0.192 0.473 0.500 0.415 0.502 0.478 0.138 0.434 0.488 0.534 0.454 0.099 0.319 0.316 0.030 0.136 0.118 0.078 0.021 0.017 0.115 0.075 0.349 0.406 0.528 0.508 0.642 0.745 0.564 0.466 0.743 0.704 0.616 0.577 0.360 0.466 0.377 0.003 0.340 0.180 0.136 0.101 0.014 0.000 0.107 ICAM Correlation CH Efflux 0.289 0.391 0.400 0.440 0.446 0.347 0.496 0.008 0.015 0.030 0.031 0.201 0.016 0.134 0.055 0.057 0.098 0.206 0.263 0.254 0.298 0.384 0.351 0.301 0.366 0.386 0.356 0.308 0.206 0.101 0.256 0.398 0.341 0.041 0.028 0.184 0.006 0.056 0.071 0.213 0.136 0.253 0.499 0.615 0.689 0.651 0.667 LDL LDL Oxidation 0.570 0.566 0.657 0.865 0.861 0.870 0.693 0.364 0.056 0.285 0.419 0.537 0.292 0.032 0.152 0.127 0.087 0.051 0.028 0.047 0.004 0.018 0.025 0.294 0.273 0.234 0.297 0.250 0.321 0.435 0.564 0.356 0.613 0.345 0.334 0.402 0.567 0.509 0.515 0.301 0.354 0.001 0.002 0.056 0.071 0.101 0.227

161

Chapter 6. Effects of type 2 diabetes on lipoprotein composition and arterial stiffness in adolescents and young adults

Introduction

Type 2 diabetes (T2D) is a major risk factor for the development of cardiovascular disease (CVD) 1 as nearly 70% of adults with T2D die of cardiovascular related complications 2. A key feature linking diabetes to CVD is the presence of an atherogenic dyslipidemia characterized by reduced levels of high density lipoprotein cholesterol (HDL-C) and increased concentrations of very low density lipoproteins

(VLDL) and small dense low density lipoproteins (LDL) 3. While LDL lowering therapy has proven effective to reduce CVD risk in adults with T2D, there is still a higher than expected residual incidence of CVD in this group 4 that may be partially explained by low levels of HDL-C 5 .

Recent therapies aimed at reducing CVD risk have focused on HDL because of its anti-atherogenic properties. HDL is best recognized for its ability to shuttle excess cholesterol from peripheral tissues to the liver for excretion in the process of “reverse cholesterol transport”, likely contributing to the well-known inverse relationship between

HDL-C and CVD in large population studies. Additional cardioprotective roles for HDL have been identified that include anti-inflammatory, anti-oxidative and anti-apoptotic properties 6.

162

While HDL-C associates clearly with CVD protection in large populations, therapies aimed at raising HDL-C have not been able to reduce CVD risk in individuals.

Specifically, the recent failure of two trails designed to raise HDL-C concentrations 7, 8 has sparked intense interest to find alternate methods to assess the cardioprotective effects of HDL. Recent studies argue that the total pool of HDL within a given individual is composed of numerous particle subpopulations with distinct protein and lipid compositions that may perform unique functions 9.

Proteomics studies consistently identify over 50 different proteins associated with

HDL 10. These finding suggests that individuals possess unique HDL “portfolios”.

These portfolios, when considered in total, may establish an individual’s risk for CVD.

HDL isolated from adults with T2D is known to carry oxidative and glycation modifications 11, particularly on its most abundant protein constituent apolipoprotein

(apo)A-I. Furthermore, HDL from individuals with diabetes has been shown to have impaired cholesterol efflux capacity 12 and reduced anti-oxidative 13 and anti-apoptotic functions 14. These changes in HDL function may be a result of poor glycemic control or oxidative stress in the setting of T2D.

Previously, our group has used ultracentrifugation and mass spectrometry to track the distribution of HDL associated proteins across particles of various densities 15.

More recently, we used gel filtration chromatography to fractionate plasma from healthy adults into 18 phospholipid containing fractions and then used a phospholipid binding resin and mass spectrometry to identify lipid associated proteins across HDL particles of various size 16 (Chapter 2). While the major apolipoproteins such as apoA-I and apoA-II were present in most fractions, 45 lower abundance HDL proteins exhibited distinct

163 distribution patterns suggestive of proteomically distinct lipoprotein subspecies in healthy male adults.

Our objective in this study was to examine the effects of early diabetes on the lipid and protein compositions of HDL in adolescents and young adults, a population that has not been studied. Additionally, we sought to relate these findings to early markers of vascular damage, carotid intimal media thickness (IMT) and pulse wave velocity (PWV), that are known to predict cardiovascular events such as myocardial infarction and stroke 17, 18.

Methods

Study Population

Post pubertal adolescents and young adults ages 16-23 years old were recruited from an established database at Cincinnati Children’s Hospital Medical Center. Only males were included in this study to eliminate known differences in plasma HDL cholesterol concentrations among races 19-21, genders 21, 22 and menstrual cycles 23.

Healthy control participants (n = 9) lacked evidence of any chronic disease and if ≤ 20 years old had a body mass index < the 85th percentile based on age and gender criteria from the Centers for Disease Control and Prevention. If they were >20 years old healthy participants had a BMI <25. Participants with T2D (n = 10) carried the diagnosis based on the American Diabetes Association criteria 24. Participants either had fasting plasma glucose ≥ 126mg/dl, a 2 hour plasma glucose >200 mg/dl during an oral glucose tolerance test, or the classic symptoms of hyperglycemia (polyuria and polydipsia) and a

164 random plasma glucose > 200 mg/dl. Individuals with type 2 diabetes were non-insulin requiring in the basal state to prevent diabetic ketoacidosis and were islet cell antibody negative indicating they did not have type 1 diabetes. None of the participants were on lipid lowering medications or had a history of smoking. Average duration of diabetes for the T2D group was 4 years 9 months.

Prior to enrollment in the study, written informed consent was obtained from participants ≥18-yr old or the parent or guardian with written assent for participants <18- yr old according to the guidelines established by the Institutional Review Board at

Cincinnati Children’s Hospital Medical Center and in accordance with the Declaration of

Helsinki.

Lipids and Lipoprotein Analyses

After a 10 hour fast, blood was drawn. One aliquot was used to obtain a fasting lipid panel. Samples were analyzed in a laboratory that is National Heart, Lung, and

Blood Institute/Centers for Disease Control and Prevention standardized with the low- density lipoprotein (LDL) cholesterol concentration calculated using the Friedewald equation. In instances where triglycerides exceeded 400mg/dl, LDL was measured by direct assay. A second aliquot was collected in a BD Vacutainer using citrate as anticoagulant. Plasma was isolated by centrifugation at 1250 x g for 15 min at 4º C and applied to three Superdex 200 gel filtration columns (GE Healthcare) arranged in series using our previously described protocol 16. Eighteen collected fractions were analyzed for phospholipid (Wako) and cholesterol (Pointe Scientific) content using colorimetric assays.

165

Proteomics

Fractions were applied to a calcium silica hydrate phospholipid binding resin

(Lipid Removal Agent (LRA); Supelco) to isolate lipoprotein particles (300 µL fraction/1.5 mg LRA). Protein components were digested with trypsin while in contact with the LRA resin and then peptide fragments were collected, reduced (dithiothreitol,

10 mM at 37º C for 30 min.) and carbamidomethylated (iodoacetamide, 40 mM at room temp. in dark for 30 min.). Peptides were dried in a Speedvac and stored at -20º C until mass spectrometry (MS) analysis. For MS analysis, peptides were resuspended in 15

µL of solvent A (95% water, 5% acetonitrile, 0.1% formic acid). Samples (0.5 µL) were injected onto a C-18 reverse phase column (Grace 218MS 150 x 0.500 mm, 5 µm) on line with and electrospray ionization time of flight mass spectrometer (QStar XL; Applied

Biosystems). Acquired mass data was scanned against the UniProtKB/Swiss-Prot

Protein Knowledgebase (release 57.0, 03/2009) using both Mascot (version 2.1) and

XTandem (version 2007.01.01.1) search engines. Positive protein identification required 90% confidence by Protein Prophet Algorithm 25. Proteomics methods are described in more detail here 16. Spectral counting was used to semi-quantitate differences in proteins across fractions between the two groups 26. We validated this method by comparing spectral counts to a quantitative biochemical technique, western blotting for the common HDL protein, apoA-I. Both techniques showed similar distribution patterns across fractions (Supplemental Figure 6-1, end of chapter). In no instance was spectral counting used to draw conclusions about the absolute quantity or concentration of a protein.

166

Arterial Thickness and Stiffness Measurements

Following the blood draw, each participant underwent noninvasive arterial vessel imaging. Arterial thickness was evaluated in the carotid arteries using high-resolution

B-mode ultrasonography (GE Vivid 7) and a 7.5 MHz linear array transducer. For each participant the right and left common carotid segments were examined independently from continuous angles to identify the thickest area of intima media thickness (IMT) from the far wall of the artery, reported as the average value in millimeters. All images were read offline using a Camtronic Medical System by a research-trained vascular technician. A trace technique was employed to measure the maximum carotid thickness from leading edge (lumen-intima) to leading edge (media-adventitia). This technique is more reproducible than point-to-point measurements with coefficients of variation for repeat trace readings of 4.1-5.3% versus 5.4% to 7.1%, respectively.

Arterial stiffness was assessed by pulse wave velocity (PWV) using a

SphygmoCor SCOR-PVx System (Atcor Medical, Sydney, Australia). The distance from a proximal artery (carotid) to distal artery recording site (femoral artery) was measured to the nearest 0.1 cm, twice, averaged and entered into the software. A tonometer was used to collect proximal and distal arterial waveforms gated by the R-wave on a simultaneously recorded electrocardiogram. PWV was then calculated as the distance from the carotid-to femoral path length divided by the time delay measured between the feet of the two waveforms reported in meters per second 17. Carotid to femoral PWV represents the aortic stiffness and is an index of arterial distensibility 17, 27. An increase in PWV has been shown to be associated with a future risk to develop coronary artery

167 disease and stroke 17. Three recordings of PWV were obtained on each participant and averaged. Repeat measures in our laboratory show coefficients of variability <7% 28.

Statistics

All data presented are means ± sample standard deviation or standard error. All comparisons were done using two-tailed student’s t-test with equal variance. A p value

< 0.05 was considered statistically significant. In general there were lower peptide counts for most of the detected proteins in the T2D group vs. the healthy individuals

(see text and Table 2). The reason for this decrease is not clear. However, to identify proteins that underwent changes in size profile in the T2D group, we devised a peptide count adjustment procedure to normalize the bulk of peptide counts across all groups.

For each protein, an adjustment factor was calculated by dividing the T2D counts by the control counts. The averaged adjustment factor (1.42) from all identified proteins was multiplied by the raw T2D counts in each fraction to derive an adjusted peptide count.

An example of this adjustment is shown in Fig. 5. Two tailed t-tests using the adjusted values were used to identify differences that were beyond those resulting from a general decrease in T2D plasma proteins noted above.

Results

Characteristics of study population

The clinical characteristics of the study population are listed in Table 6-1. The groups were similar in age with no significant differences in blood pressure. Compared to healthy adolescents and young adults, youth with T2D had an increased body mass

168 index (BMI) and higher total cholesterol (p<0.05). There was no statistically significant difference in LDL cholesterol levels between the groups but HDL cholesterol was decreased in the T2D group (p<0.001) which has been observed in other studies 28, 29.

Significant variability in the triglyceride measurements in the T2D group obscured a statistically significant difference from the healthy individuals, though it is clear that the

T2D group trended toward higher triglyceride levels. PWV was higher in the T2D group demonstrating increased arterial stiffness compared to healthy youth (p<0.05) similar to that shown in larger studies in youth 28. There were no differences in the common carotid IMT measurements between the two groups. Lack of carotid differences may be due to a small number of participants or the fact that changes in carotid IMT may reflect more advanced arterial pathology compared to changes in PWV 30.

Lipoprotein profiles

The phospholipid and cholesterol distributions across the 18 lipid containing plasma fractions were determined by enzymatic assay and are presented in Figure 6-1A and B. In this technique, LDL/VLDL sized particles migrate together in Peak 1 (fractions

14-18) and HDL particles are broadly distributed across Peak 2 (fractions 19-30) as the technique has been optimized to spread particles in the HDL size range 16. Compared to healthy youth, those with T2D had higher phospholipid content in LDL/VLDL containing fractions 16 and 17 (p<0.05). T2D youth also exhibited lower phospholipid content in the larger HDL containing fractions 21 - 24 (p<0.01) compared to controls.

There were no phospholipid differences in the smaller HDL particles (fractions 25- 29).

169

Cholesterol content was higher in LDL/VLDL fractions 15 - 17 in the T2D group

(p<0.05), mirroring the phospholipid content. Interestingly, cholesterol associated with smaller HDL fractions 24 – 27 was increased in T2D compared to healthy youth

(p<0.01). This was surprising given that the clinical measure (Table 6-1) indicated lower

HDL-C in the T2D group. The reason for this disparity is not clear but likely involves inherent differences between the gel filtration separation and the clinical apoB precipitation methodologies for assessment of HDL-C.

Taken together, the phospholipid and cholesterol distributions indicate that individuals with T2D generally have smaller HDL particles that are enriched in cholesterol relative to phospholipid, while healthy adolescents exhibit a population of larger, more phospholipid-rich HDL particles exhibited by fractions 21-24.

Clinical lipid measurements and arterial stiffness measurements

We first determined if commonly used clinical lipid measures were associated with arterial stiffness measurements in this cohort. The univariate correlations between

PWV, HDL-C, LDL-C and total cholesterol (TC) are shown in Figure 6-2. While a significant correlation was found between total plasma cholesterol and PWV (panel C), we found no clear association between the clinical lipid measures of HDL-C or LDL-C and PWV (panels A and B). Similar analyses showed no associations with common carotid IMT.

Lipoprotein profiles and arterial stiffness measurements

170

Next, we sought to determine if lipoprotein size subfractions from our gel filtration analysis were more tightly associated with arterial stiffness as measured by PWV.

Associations with common carotid IMT were not assessed due to the lack of differences between groups. Figure 6-3 shows the correlation coefficients of PWV vs. lipid content for each individual fraction with statistically significant differences between the healthy and T2D groups indicated with a black bar (p<0.05). Combining all patients in the study, Fig. 6-3A shows that the phospholipid content of LDL/VLDL fractions 14, 15, and

16 and smaller HDL fractions 25, 26 and 27 were significantly and positively correlated with PWV, while larger HDL fractions 21, 22, and 23 were negatively correlated (p<

0.05). Similar relationships were identified for the cholesterol content across the fractions, though the correlations with phospholipid content were stronger (Fig. 6-3D).

When the same analysis was performed on data from the healthy participants only, a similar pattern of correlations was apparent for both phospholipid and cholesterol (Fig.

6-3B, E). However, in the T2D group, (Fig. 6-3C, F) the correlations in the large phospholipid-rich HDL fractions 21-24 disappear, because these patients effectively lack these particles.

Figure 6-4 further illustrates the relationships between phospholipid content and

PWV. The univariate correlations are shown for the most significant fraction in each of the regions identified in Fig. 6-3A. The phospholipid concentrations among the healthy youth and the T2D group overlap significantly in the VLDL/LDL (fraction 16, Fig. 6-4A) and smaller HDL fractions (fraction 26, Fig. 6-4C). However, in the large HDL fraction

(fraction 22, Fig. 6-4B) there is a well-defined breakpoint between groups with all T2D participants exhibiting less than 0.01 µg/µl phospholipid in this fraction. By contrast, all

171 healthy individuals were above this threshold. In summary, the above data indicate that

1) the phospholipid content of fractions better predict increased arterial stiffness compared to the clinical lipid measurements, and 2) the strongest association appears to be the inverse relationship between larger HDL fractions and PWV.

Proteomics

We next set out to determine the proteomic make-up of the fractionated lipoproteins reported above. Across all samples and all participants, we identified 45 lipid associated proteins (Table 6-2) using our previously published criteria 16. Nearly all proteins were detectable in both groups with only two that were unique to the healthy group, serum amyloid A and serotransferrin. Table 6-2 shows that, compared to healthy participants, youth with T2D exhibited decreased spectral counts for most of the identified proteins (when summed across all fractions), including the major HDL protein apoA-I. This is consistent with the overall lower amount of phospholipid that we detected in the HDL fractions of T2D vs. controls (Fig. 6-1A) and the lower plasma HDL-

C measured in the clinical assay (Table 6-1).

The fact that the T2D participants exhibited overall lower spectral counts for most

HDL proteins complicated the task of determining whether specific subfractions are altered in the face of T2D. To circumvent this problem, we normalized the spectral counts from the T2D group based on an adjustment factor that was derived from a global comparison of spectral counts from all HDL proteins between fractions 19-30

(see methods). We then compared the corrected T2D data to the healthy youth and examined the profiles that were different even after the global correction. Fig. 6-5A

172 shows the column proteomic profile for apoA-I prior to adjustment. Looking across all fractions, it is clear that the T2D exhibited about 40% lower apoA-I levels than the controls. Fig. 6-5B shows the comparison after correction indicating a largely similar profile between the two groups, though differences remained in the larger HDL particles.

Figure 6-6A-C shows several examples where proteins exhibited clear differences between the T2D and healthy groups while Figure 6-6D-F shows fraction distribution profiles for some examples of proteins that were clearly similar between the two groups.

One of the most striking differences was apoE, which was found in larger HDL species in healthy individuals, but its spectral counts were reduced nearly 5-fold and shifted toward smaller HDL particles in the T2D group. ApoC-I and paraoxonase (PON)-1 also showed an overall lower abundance in the T2D group. There were no clear differences within the LDL/VLDL containing fractions among these proteins (data not shown).

However, as mentioned, our techniques for detecting low abundance proteins on LDL and VLDL have not been optimized.

Table 6-3 shows the p-values resulting from a t-test comparison of the total peptide counts measured in healthy participants vs. the adjusted values from the T2D group. Statistically significant differences are noted with red highlighting. Across the various fractions we noted significant decreases in 17 proteins in the T2D vs. the healthy group. There were no instances of increased levels in the T2D group.

Interestingly, a majority of these changes occurred in fractions 21-23, the same fractions that exhibited significant reductions in phospholipid in the T2D group (and the ones inversely correlated with PWV).

173

Discussion

The hypothesis driving this study states that specific lipoprotein subpopulations, beyond the traditional density-centric distinctions of “LDL” or “HDL”, may differentially participate in arterial disease development in the setting of T2D. If this idea is correct, then identification of specific altered populations could lead to better biomarkers for assessing diabetic disease progression, perhaps in its earliest stages. Furthermore, if unique particles could be identified, this could yield important pathways for therapeutic exploitation. In this study, we attempted to correlate lipoprotein subpopulations with early indicators of vascular disease. Our logic was that changes in adolescents with

T2D may better identify lipoproteins that are altered early in the disease course.

Capitalizing on our previous work showing that fresh human plasma can be rapidly and natively separated into an array of particles with distinct protein profiles 16, we have now correlated these fractions with early indications of vascular dysfunction in T2D adolescents and young adults.

Before discussing the implications of our findings, it is worthwhile to clarify some issues surrounding lipoprotein nomenclature. Lipoproteins are most commonly isolated by density ultracentrifugation and the major classes have been historically defined

(indeed named) by their density 31. In contrast, we elected to analyze lipoproteins by size through gel filtration chromatography because we and others have noted fewer alterations in particle proteome compared to the harsher conditions inherent to centrifugation 9, 16. Although lipoprotein size and density are roughly inversely correlated, there are important exceptions to this rule. For example, trypanosome lytic

174 factor 2 is a protein-rich particle containing immunoglobulin M, apoA-I and haptoglobin related protein with lipid making up less than 1% of its mass. By ultracentrifugation, this particle would migrate among the densest HDL particles. However, its large size results in co-migration with LDL sized particles by gel filtration 32. In order to relate our gel filtration results to traditional density-centric definitions, we elected to use the presence of apoB, the core constituent of LDL, as the key distinguisher. Therefore, the

“VLDL/LDL” range is defined as fractions 14 - 18 due to the presence of apoB. We assigned the remaining fractions 19 - 30 as the "HDL" range because their diameters are consistent with measurements for density isolated HDL and because of the abundance of the major HDL protein apoA-I. However, we caution that it is quite possible that particles exist within these “HDL” fractions that may be outside the density range of 1.063-1.210 g/ml or may not contain apoA-I.

In this study, we did not note correlations between HDL-C and LDL-C and PWV.

Prior work in a larger cohort of T2D adolescents (n=436) also failed to demonstrate a relationship between these lipid measures and PWV 28. Lack of associations between the clinical lipid measures and increased PWV is likely because these clinical measures are not robust enough to predict CVD risk in individuals. However, our results clearly identify specific size populations within both the LDL and HDL size ranges that correlate tightly with early vascular stiffness. This association is true despite the relatively small number of participants in this study. For example, fractions 14-16, within the VLDL/LDL size range, positively correlated with PWV, tracked either by phospholipid or cholesterol content (Fig. 6-3) while the clinical LDL-C measure failed to correlate (Fig 6-2B). In the

HDL size range, we noted two regions that tightly associated with PWV. Fractions 21-

175

23, representing larger particles, were strongly inversely correlated (i.e. associated with protection) while fractions 25-27, representing smaller particles, were strongly positively correlated (i.e. associated with pathology). Therefore, depending on the subpopulation distribution in a given individual, it is possible that the balance between these fractions may result in greater or less atheroprotection.

The fact that different particles within the HDL range show widely different correlations with vascular stiffness supports the idea that specific HDL subpopulations are differentially associated with protection or pathology in the vessel wall. We were particularly intrigued by the fact that all of the T2D participants exhibited less than 0.01

µg/µl of phospholipid in fraction 22, whereas none of the healthy individuals were below this threshold. Fraction 22 was the only peak identified that could definitively distinguish between T2D and healthy participants. Thus, particles within fraction 22 may prove useful as biomarkers for early onset of vascular disease in the setting of T2D. In addition, documenting temporal changes in these fractions over the course of disease and changes in medical therapy may prove beneficial.

Our proteomics analyses were consistent with our previous studies 16 in healthy adult males showing that the various HDL-associated proteins distribute in distinct patterns across the size gradient (Supplemental Figure 6-2A-B). After applying a correction for dealing with the overall reduced levels of most proteins detected in the

T2D group, we identified several proteins that undergo additional abundance or size distribution changes in the setting of T2D. Interestingly, in the case of the particles in the HDL size range, most of these changes corresponded with the fractions that were negatively associated with PWV. In fractions 21-23, 9 proteins exhibited reductions in

176 the T2D group. One of the most striking was apoE which underwent a nearly 5-fold reduction in peptide count abundance in these fractions.

ApoE is widely distributed on VLDL, LDL and HDL in human plasma and is known to associate with larger HDL particles. In fact, much of apoE may exist on HDL particles that lack apoA-I 33, 34. Genetic ablation of apoE in mice is well known to cause a variety of vascular phenotypes including atherosclerosis 35, endothelial dysfunction 36, and alterations vascular remodeling and restenosis 37. However, the role of HDL- associated apoE on vascular function has not been as widely characterized, though it has been suggested that apoE-containing HDL stimulates the production of atheroprotective heparin sulfate proteoglycans 38. Another protein that exhibited impressive decreases in the HDL particle size range was paraoxonase 1. This is a serum esterase that hydrolyzes organophosphates and has been shown by numerous studies to play an important role in the prevention and scavenging of lipid peroxidation in LDL 39. Paraoxonase 1 has been implicated in the severity of atherosclerosis in T2D adults 40 and may protect against diabetes via its antioxidative effects and/or through effects on insulin secretion 41. Its activity is reduced in the setting of T2D, though this may be due to reductions in specific activity rather than lower serum concentrations 42.

Low paraoxonase activities have also been linked to arterial stiffness as measured by carotid-femoral PWF in hypertensive 43 and renal transplant patients 44. ApoM was also lower in the larger HDL sized particles in the T2D group, consistent with whole plasma reductions of around 10% noted by others 45. This HDL-associated lipocalin does not have a well-defined function, but its transgenic over expression in mice leads to reduced atherosclerotic lesions, possibly by facilitating HDL remodeling, increased cholesterol

177 efflux 46, or its effects on the generation of particularly atheroprotective pre-beta forms of

HDL 45.

With regard to the smaller fractions of HDL (25-27) that were positively associated with PWV, we noted relatively few changes in protein abundance in these fractions and it is not obvious how reductions in these proteins might affect vascular function. We did find it interesting that these particles tended to be highly cholesterol enriched relative to phospholipid in the T2D population. One possibility is that these cholesterol laden particles are dysfunctional in one or more of the known atheroprotective characteristics of HDL.

While it is tempting to speculate that the lipoprotein size fractions identified in this study may be directly involved in either vascular protection or pathology, we caution that our data is purely correlative at this point. We do not know if the identified particle populations are causative or are by-products resulting from the action of more directly relevant pathways. Nevertheless, the strong correlations of the size subpopulations with arterial stiffness are clear and we believe that our data strongly justify efforts to biochemically identify the specific particles that are changing in early stages of diabetes.

While a gel filtration analysis like we used here will likely never be practical in a clinical setting, this information could be highly useful for developing clinical assays for the early detection of vascular pathology and for measuring the efficacy of therapeutic or dietary interventions.

178

Reference List

1 Grundy SM, Garber A, Goldberg R, Havas S, Holman R, Lamendola C, Howard

WJ, Savage P, Sowers J, Vega GL. Prevention Conference VI: Diabetes and

Cardiovascular Disease: Writing Group IV: lifestyle and medical management of

risk factors. Circulation 2002;105(18):e153-e158.

2 Panzram G. Mortality and survival in type 2 (non-insulin-dependent) diabetes

mellitus. Diabetologia 1987;30(3):123-31.

3 Adiels M, Olofsson SO, Taskinen MR, Boren J. Diabetic dyslipidaemia. Curr Opin

Lipidol 2006;17(3):238-46.

4 Fruchart JC, Sacks FM, Hermans MP, Assmann G, Brown WV, Ceska R,

Chapman MJ, Dodson PM, Fioretto P, Ginsberg HN, Kadowaki T, Lablanche JM,

Marx N, Plutzky J, Reiner Z, Rosenson RS, Staels B, Stock JK, Sy R, Wanner C,

Zambon A, Zimmet P. The Residual Risk Reduction Initiative: a call to action to

reduce residual vascular risk in dyslipidaemic patient. Diab Vasc Dis Res

2008;5(4):319-35.

5 Barter P, Gotto AM, LaRosa JC, Maroni J, Szarek M, Grundy SM, Kastelein JJ,

Bittner V, Fruchart JC. HDL cholesterol, very low levels of LDL cholesterol, and

cardiovascular events. N Engl J Med 2007;357(13):1301-10.

6 Gordon SM, Hofmann S, Askew DS, Davidson WS. High density lipoprotein: it's

not just about lipid transport anymore. Trends Endocrinol Metab 2011;22(1):9-15.

7 Barter PJ, Caulfield M, Eriksson M, Grundy SM, Kastelein JJ, Komajda M, Lopez-

Sendon J, Mosca L, Tardif JC, Waters DD, Shear CL, Revkin JH, Buhr KA, Fisher

179

MR, Tall AR, Brewer B. Effects of torcetrapib in patients at high risk for coronary

events. N Engl J Med 2007;357(21):2109-22.

8 Schwartz GG, Olsson AG, Ballantyne CM, Barter PJ, Holme IM, Kallend D, Leiter

LA, Leitersdorf E, McMurray JJ, Shah PK, Tardif JC, Chaitman BR, Duttlinger-

Maddux R, Mathieson J. Rationale and design of the dal-OUTCOMES trial:

efficacy and safety of dalcetrapib in patients with recent acute coronary syndrome.

Am Heart J 2009;158(6):896-901.

9 Gordon S, Durairaj A, Lu J, Davidson WS. High-density lipoprotein proteomics:

Identifying new drug targets and biomarkers by understanding functionality. Curr

Cardio Risk Rep 2010;4(Volume 4, Number 1):1-8.

10 Davidson WS. Davidson Lab Webpage: HDL Proteome Watch.

http://homepages.uc.edu/~davidswm/HDLProteomeList.xlsx . 2012.

11 Kontush A, Chapman MJ. Why is HDL functionally deficient in type 2 diabetes?

Curr Diab Rep 2008;8(1):51-9.

12 Syvanne M, Castro G, Dengremont C, de Geitere C, Jauhiainen M, Ehnholm C,

Michelagnoli S, Franceschini G, Kahri J, Taskinen MR. Cholesterol efflux from

Fu5AH hepatoma cells induced by plasma of subjects with or without coronary

artery disease and non- insulin-dependent diabetes: importance of LpA-I:A-II

particles and phospholipid transfer protein. Atherosclerosis 1996;127(2):245-53.

13 Nobecourt E, Jacqueminet S, Hansel B, Chantepie S, Grimaldi A, Chapman MJ,

Kontush A. Defective antioxidative activity of small dense HDL3 particles in type 2

diabetes: relationship to elevated oxidative stress and hyperglycaemia.

Diabetologia 2005;48(3):529-38.

180

14 de Souza JA, Vindis C, Hansel B, Negre-Salvayre A, Therond P, Serrano CV, Jr.,

Chantepie S, Salvayre R, Bruckert E, Chapman MJ, Kontush A. Metabolic

syndrome features small, apolipoprotein A-I-poor, triglyceride-rich HDL3 particles

with defective anti-apoptotic activity. Atherosclerosis 2008;197(1):84-94.

15 Davidson WS, Silva RA, Chantepie S, Lagor WR, Chapman MJ, Kontush A.

Proteomic analysis of defined HDL subpopulations reveals particle-specific protein

clusters: relevance to antioxidative function. Arterioscler Thromb Vasc Biol

2009;29(6):870-6.

16 Gordon SM, Deng J, Lu LJ, Davidson WS. Proteomic characterization of human

plasma high density lipoprotein fractionated by gel filtration chromatography. J

Proteome Res 2010;9:5239-49.

17 Laurent S, Cockcroft J, Van BL, Boutouyrie P, Giannattasio C, Hayoz D, Pannier

B, Vlachopoulos C, Wilkinson I, Struijker-Boudier H. Expert consensus document

on arterial stiffness: methodological issues and clinical applications. Eur Heart J

2006;27(21):2588-605.

18 Bots ML, Hoes AW, Koudstaal PJ, Hofman A, Grobbee DE. Common carotid

intima-media thickness and risk of stroke and myocardial infarction: the Rotterdam

Study. Circulation 1997;96(5):1432-7.

19 Goldberg AP, Harter HR, Patsch W, Schechtman KB, Province M, Weerts C, Kuisk

I, McCrate MM, Schonfeld G. Racial differences in plasma high-density lipoproteins

in patients receiving hemodialysis. A possible mechanism for accelerated

atherosclerosis in white men. N Engl J Med 1983;308(21):1245-52.

181

20 Heiss G, Schonfeld G, Johnson JL, Heyden S, Hames CG, Tyroler HA. Black-white

differences in plasma levels of apolipoproteins: the Evans County Heart Study. Am

Heart J 1984;108(3 Pt 2):807-14.

21 Johnson JL, Slentz CA, Duscha BD, Samsa GP, McCartney JS, Houmard JA,

Kraus WE. Gender and racial differences in lipoprotein subclass distributions: the

STRRIDE study. Atherosclerosis 2004;176(2):371-7.

22 Pascot A, Lemieux I, Bergeron J, Tremblay A, Nadeau A, Prud'homme D, Couillard

C, Lamarche B, Despres JP. HDL particle size: a marker of the gender difference

in the metabolic risk profile. Atherosclerosis 2002;160(2):399-406.

23 Muesing RA, Forman MR, Graubard BI, Beecher GR, Lanza E, McAdam PA,

Campbell WS, Olson BR. Cyclic changes in lipoprotein and apolipoprotein levels

during the menstrual cycle in healthy premenopausal women on a controlled diet. J

Clin Endocrinol Metab 1996;81(10):3599-603.

24 Standards of medical care in diabetes--2012. Diabetes Care 2012;35 Suppl 1:S11-

S63.

25 Nesvizhskii AI, Keller A, Kolker E, Aebersold R. A statistical model for identifying

proteins by tandem mass spectrometry. Anal Chem 2003;75(17):4646-58.

26 Vaisar T, Pennathur S, Green PS, Gharib SA, Hoofnagle AN, Cheung MC, Byun J,

Vuletic S, Kassim S, Singh P, Chea H, Knopp RH, Brunzell J, Geary R, Chait A,

Zhao XQ, Elkon K, Marcovina S, Ridker P, Oram JF, Heinecke JW. Shotgun

proteomics implicates protease inhibition and complement activation in the

antiinflammatory properties of HDL. J Clin Invest 2007;117(3):746-56.

182

27 Asmar R, Benetos A, Topouchian J, Laurent P, Pannier B, Brisac AM, Target R,

Levy BI. Assessment of arterial distensibility by automatic pulse wave velocity

measurement. Validation and clinical application studies. Hypertension

1995;26(3):485-90.

28 Urbina EM, Kimball TR, Khoury PR, Daniels SR, Dolan LM. Increased arterial

stiffness is found in adolescents with obesity or obesity-related type 2 diabetes

mellitus. J Hypertens 2010;28(8):1692-8.

29 Urbina EM, Kimball TR, McCoy CE, Khoury PR, Daniels SR, Dolan LM. Youth with

obesity and obesity-related type 2 diabetes mellitus demonstrate abnormalities in

carotid structure and function. Circulation 2009;119(22):2913-9.

30 Koivistoinen T, Virtanen M, Hutri-Kahonen N, Lehtimaki T, Jula A, Juonala M,

Moilanen L, Aatola H, Hyttinen J, Viikari JS, Raitakari OT, Kahonen M. Arterial

pulse wave velocity in relation to carotid intima-media thickness, brachial flow-

mediated dilation and carotid artery distensibility: the Cardiovascular Risk in Young

Finns Study and the Health 2000 Survey. Atherosclerosis 2012;220(2):387-93.

31 Lindgren FT, Elliot HA, Gofman JW. The ultracentrifugal characterization and

isolation of human blood lipids and lipoproteins, with applications to the study of

atherosclerosis. J Phys Colloid Chem 1951;55(1):80-93.

32 Terasaka N, Westerterp M, Koetsveld J, Fernandez-Hernando C, Yvan-Charvet L,

Wang N, Sessa WC, Tall AR. ATP-Binding Cassette Transporter G1 and High-

Density Lipoprotein Promote Endothelial NO Synthesis Through a Decrease in the

Interaction of Caveolin-1 and Endothelial NO Synthase. Arterioscler Thromb Vasc

Biol 2010;30(11):2219-25.

183

33 Gibson JC, Rubinstein A, Brown WV, Ginsberg HN, Greten H, Norum R, Kayden

H. Apo E-containing lipoproteins in low or high density lipoprotein deficiency.

Arteriosclerosis 1985;5(4):371-80.

34 Santos RD, Schaefer EJ, Asztalos BF, Polisecki E, Wang J, Hegele RA, Martinez

LR, Miname MH, Rochitte CE, Da Luz PL, Maranhao RC. Characterization of high

density lipoprotein particles in familial apolipoprotein A-I deficiency. J Lipid Res

2008;49(2):349-57.

35 Zhang SH, Reddick RL, Piedrahita JA, Maeda N. Spontaneous

hypercholesterolemia and arterial lesions in mice lacking apolipoprotein E. Science

1992;258(5081):468-71.

36 Yang R, Powell-Braxton L, Ogaoawara AK, Dybdal N, Bunting S, Ohneda O, Jin H.

Hypertension and endothelial dysfunction in apolipoprotein E knockout mice.

Arterioscler Thromb Vasc Biol 1999;19(11):2762-8.

37 Ali ZA, Alp NJ, Lupton H, Arnold N, Bannister T, Hu Y, Mussa S, Wheatcroft M,

Greaves DR, Gunn J, Channon KM. Increased in-stent stenosis in ApoE knockout

mice: insights from a novel mouse model of balloon angioplasty and stenting.

Arterioscler Thromb Vasc Biol 2007;27(4):833-40.

38 Paka L, Kako Y, Obunike JC, Pillarisetti S. Apolipoprotein E containing high

density lipoprotein stimulates endothelial production of heparan sulfate rich in

biologically active heparin-like domains. A potential mechanism for the anti-

atherogenic actions of vascular apolipoprotein e. J Biol Chem 1999;274(8):4816-

23.

184

39 Mackness MI, Arrol S, Durrington PN. Paraoxonase prevents accumulation of

lipoperoxides in low-density lipoprotein. FEBS Lett 1991;286(1-2):152-4.

40 Tartan Z, Orhan G, Kasikcioglu H, Uyarel H, Unal S, Ozer N, Ozay B, Ciloglu F,

Cam N. The role of paraoxonase (PON) enzyme in the extent and severity of the

coronary artery disease in type-2 diabetic patients. Heart Vessels 2007;22(3):158-

64.

41 Koren-Gluzer M, Aviram M, Meilin E, Hayek T. The antioxidant HDL-associated

paraoxonase-1 (PON1) attenuates diabetes development and stimulates beta-cell

insulin release. Atherosclerosis 2011;219(2):510-8.

42 Abbott CA, Mackness MI, Kumar S, Boulton AJ, Durrington PN. Serum

paraoxonase activity, concentration, and phenotype distribution in diabetes mellitus

and its relationship to serum lipids and lipoproteins. Arterioscler Thromb Vasc Biol

1995;15(11):1812-8.

43 Yang WI, Lee SH, Ko YG, Kang SM, Choi D, Ha JW, Hong MK, Chung N, Shim

WH, Cho SY, Jang Y. Relationship between paraoxonase-1 activity, carotid intima-

media thickness and arterial stiffness in hypertensive patients. J Hum Hypertens

2010;24(7):492-4.

44 Gungor O, Kircelli F, Demirci MS, Tuncel P, Sisman AR, Tatar E, Hur E, Asci G,

Ok E, Toz H. Serum paraoxonase 1 activity predicts arterial stiffness in renal

transplant recipients. J Atheroscler Thromb 2011;18(10):901-5.

45 Plomgaard P, Dullaart RP, de VR, Groen AK, Dahlback B, Nielsen LB.

Apolipoprotein M predicts pre-beta-HDL formation: studies in type 2 diabetic and

nondiabetic subjects. J Intern Med 2009;266(3):258-67.

185

46 Christoffersen C, Jauhiainen M, Moser M, Porse B, Ehnholm C, Boesl M, Dahlback

B, Nielsen LB. Effect of apolipoprotein M on high density lipoprotein metabolism

and atherosclerosis in low density lipoprotein receptor knock-out mice. J Biol Chem

2008;283(4):1839-47.

186

Table 6-1. Clinical characteristics of study participants. Table 1. Clinical characteristics of study participants. Healthy Type 2 diabetes p n 9 10

Age (years) 21.8 ± 2.5 21.5 ± 2.9 0.827

Body mass index (mg/k2) 23.8 ± 1.9 37.8 ± 6.0* <0.001

Total cholesterol (mg/dl) 157 ± 32 211 ± 51* 0.014

Triglycerides (mg/dl) 78 ± 29 286 ± 317 0.068

HDL cholesterol (mg/dl) 50.0 ± 9 32.6 ± 6* <0.001

LDL cholesterol (mg/dl) 90 ± 34 124 ± 46 0.087

Systolic BP (mmHg) 116 ± 11 119 ± 9 0.564

Diastolic BP (mmHg) 72 ± 7 73 ± 5 0.877

Pulse wave velocity (m/s) 5.81 ± 0.98 7.15 ± 1.05* 0.012

Carotid IMT (mm) 0.48 ± 0.08 0.53 ± 0.09 0.184 HDL, high density lipoprotein; LDL, low density lipoprotein; BP, blood pressure; IMT, intima media thickness. * indicates p<0.05 compared to healthy group.

187

Figure 6-1. Lipid distribution profiles are altered in type 2 diabetes. Panel A displays the phospholipid content across the multiple size based plasma fractions in healthy youth (blue) and those with type 2 diabetes (red). Panel B displays the cholesterol content across the multiple size based plasma fractions for the two groups. # indicates p < 0.05; * indicates p < 0.01.

188

Figure 6-2. Correlations of standard clinical lipid measures with pulse wave velocity. Scatter plots displaying the relationships between clinical lipid measures of HDL cholesterol (panel A), LDL cholesterol (panel B) or total cholesterol (panel C) versus pulse wave velocity. Red= type 2 diabetes; Blue= healthy189 participants.

Figure 6-3. Lipoprotein fractions correlate with pulse wave velocity. Panel A displays the correlation value (r) between the phospholipid content in each fraction and pulse wave velocity for all participants (black); B for healthy participants (blue); C for participants with type 2 diabetes (red). Panel D displays the correlation value (r) between the cholesterol content in each fraction and pulse wave velocity for all participants (black); E for healthy participants (blue); F for participants with type 2 diabetes (red). Black bars indicate fractions with p<0.05.

190

Figure 6-4. Scatterplots for the fractions with the strongest observed correlation with pulse wave velocity. Fraction 16 (A), Fraction 22 (B) and Fraction 26 (C). Red= type 2 diabetes; blue= healthy participants.

191

Table 6-2. Comparison of phospholipid associated proteins between groups. Avg. Spectral Counts Protein Count Healthy T2DM Alpha-1-antitrypsin 9 3.7 0.8 Alpha-1B-glycoprotein 13 4.7 1.8 Alpha-2-antiplasmin 11 2.4 0.5 Alpha-2-HS-glycoprotein 14 17.3 4.7 AMBP protein 12 2 2.2 Angiotensinogen 9 2.7 0.5 Antithrombin-III 13 11.6 6.2 Apolipoprotein A-I 15 145 86.2 Apolipoprotein A-II 15 13.3 6.3 Apolipoprotein A-IV 15 58.1 27 Apolipoprotein B-100 13 36.7 6.5 Apolipoprotein C-I 14 8.2 3 Apolipoprotein C-II 13 1.6 0.8 Apolipoprotein C-III 14 8.6 5.3 Apolipoprotein E 13 22 9.8 Apolipoprotein M 13 4 1.2 Apolipoprotein-L1 10 2 1.8 Beta-2-glycoprotein 1 15 9.6 4.8 C4b-binding protein alpha chain 13 7.7 2.8 Clusterin 15 23.9 15.7 Complement C1s subcomponent 13 6.7 2.3 Complement C3 15 131.7 80.5 Complement C4-B 15 74.3 35.7 Complement factor B 15 35.9 19.5 Fibrinogen alpha chain 15 54.9 22.7 Haptoglobin-related protein 12 10 4.7 Hemopexin 15 18.9 11.2 Heparin cofactor 2 15 14.1 10.8 Insulin-like growth factor-binding protein ALS 12 10.7 4.3 Inter-alpha-trypsin inhibitor heavy chain H2 15 24.3 10 Inter-alpha-trypsin inhibitor heavy chain H4 15 31 8 Kininogen-1 14 13 7.7 Lipopolysaccharide-binding protein 6 1.1 0.5 Pigment epithelium-derived factor 15 15.4 8.2 Plasma protease C1 inhibitor 7 2.3 0.5 Prothrombin 15 15.3 9 Retinol-binding protein 4 9 2.3 0.3 Serotransferrin 3 2.8 0 Serum albumin 15 142.6 70.3 Serum amyloid A 3 1.2 0 Serum amyloid A-4 7 1.4 0.3 Serum paraoxonase/arylesterase 1 12 15.8 2.7 Transthyretin 9 4.1 1.2 Vitamin D-binding protein 15 14 3.2 Vitronectin 15 16.7 11.7 Restricted to only proteins found in 3 prior HDL studies. Count is the number of subjects with > 0 peptides for given protein.

192

Figure 6-5. Adjustment of proteomics data. Distribution profiles of apoA-I across the collected fractions are shown for healthy (blue) and type 2 participants (red): A, before adjustment; B, after adjustment.

193

Figure 6-6. Distribution patterns of phospholipid associated proteins are altered in type 2 diabetes. Mass spectrometry was used to determine the relative abundance of phospholipid associated proteins across collected plasma fractions. Distribution profiles of proteins across the collected fractions are shown for healthy (blue) and type 2 participants (red); A, apolipoprotein E; B, apolipoprotein C-I; C, paraoxonase 1; D, apolipoprotein H; E, apolipoprotein J (Clusterin); F, complement C3.

194

Table 6-3. Differences in protein distributions patterns between groups. Fraction number Protein Accession 19 20 21 22 23 24 25 26 27 28 29 30 Alpha-1-antitrypsin A1AT 0.43 0.43 0.36 0.15 0.29 0.27 Alpha-1B-glycoprotein A1BG 0.39 0.23 0.83 Alpha-2-antiplasmin A2AP 0.02 0.40 0.13 Alpha-2-HS-glycoprotein FETUA 0.24 0.11 0.00 0.03 0.06 AMBP protein AMBP 0.83 0.46 0.17 0.66 Angiotensinogen ANGT 0.23 0.23 0.06 0.06 0.04 0.15 Antithrombin-III ANT3 0.22 0.55 0.74 0.78 Apolipoprotein A-I APOA1 0.30 0.58 0.02 0.02 0.01 0.14 0.74 0.77 0.47 0.74 0.48 0.36 Apolipoprotein A-II APOA2 0.23 0.01 0.00 0.05 0.95 0.15 0.72 0.24 0.56 Apolipoprotein A-IV APOA4 0.43 0.09 0.23 0.33 0.08 0.47 0.18 0.15 0.19 0.42 Apolipoprotein B-100 APOB Apolipoprotein C-I APOC1 0.43 0.24 0.16 0.09 0.03 0.32 0.70 0.55 0.06 0.24 0.43 Apolipoprotein C-II APOC2 0.60 0.06 0.73 0.24 0.43 Apolipoprotein C-III APOC3 0.93 0.80 0.24 0.89 0.94 0.90 0.78 0.38 0.24 0.43 0.23 Apolipoprotein E APOE 0.09 0.02 0.01 0.03 0.22 0.85 0.60 0.23 0.39 0.24 0.23 Apolipoprotein M APOM 0.27 0.02 0.42 0.24 0.60 Apolipoprotein-L1 APOL1 0.89 0.28 0.60 0.24 0.97 0.23 Beta-2-glycoprotein 1 APOH 0.23 0.12 0.47 0.25 0.19 C4b-binding protein alpha chain C4BPA 0.43 Clusterin (apoJ) CLUS 0.80 0.09 0.89 0.43 0.86 0.34 0.66 0.88 0.53 0.24 Complement C1s subcomponent C1S 0.09 0.05 0.80 0.12 Complement C3 CO3 0.43 0.70 0.29 0.86 0.46 0.06 0.43 0.43 Complement C4-B CO4B 0.43 0.33 0.06 0.21 0.28 0.16 0.29 Complement factor B CFAB 0.09 0.27 0.78 0.11 0.43 Fibrinogen alpha chain FIBA 0.38 0.14 0.73 0.43 0.43 0.43 0.43 0.24 Haptoglobin-related protein HPTR 0.43 0.09 0.36 0.43 0.23 Hemopexin HEMO 0.88 0.41 0.80 0.51 0.06 0.43 Heparin cofactor 2 HEP2 0.97 0.44 0.84 0.64 0.02 Insulin-like growth factor-binding protein ALS ALS 0.43 0.36 0.30 0.15 0.43 Inter-alpha-trypsin inhibitor heavy chain H2 ITIH2 0.80 0.21 0.01 0.00 0.24 Inter-alpha-trypsin inhibitor heavy chain H4 ITIH4 0.43 0.43 0.04 0.01 0.03 0.01 0.13 0.43 0.24 Kininogen-1 KNG1 0.94 0.13 0.24 0.73 0.91 0.27 0.28 0.63 0.43 Lipopolysaccharide-binding protein LBP 0.83 0.57 0.43 Pigment epithelium-derived factor PEDF 0.27 0.08 0.98 0.34 Plasma protease C1 inhibitor IC1 0.18 0.26 Prothrombin THRB 0.69 0.62 0.24 0.06 0.43 0.43 Retinol-binding protein 4 RET4 0.38 0.04 0.09 Serotransferrin TRFE 0.16 0.31 0.13 Serum albumin ALBU 0.43 0.20 0.07 0.24 0.11 0.71 0.06 0.03 0.21 0.20 0.15 0.21 Serum amyloid A protein SAA 0.43 0.16 0.16 0.24 Serum amyloid A-4 protein SAA4 0.43 0.78 0.38 0.13 0.43 Serum paraoxonase/arylesterase 1 PON1 0.27 0.02 0.00 0.04 0.06 0.01 0.15 0.15 0.17 0.43 Transthyretin TTHY 0.43 0.43 0.33 0.22 0.29 0.42 Vitamin D-binding protein VTDB 0.27 0.43 0.24 0.03 0.05 0.23 Vitronectin VTNC 0.24 0.24 0.68 0.52 0.95 0.69 0.02 19 20 21 22 23 24 25 26 27 28 29 30 Significant differences/fraction 0 0 3 6 8 2 2 3 1 3 4 0 Shaded values indicate p value <0.05. The fractions between the two vertical dotted lines were found to have a significant inverse correlation with measures of vascular health.

195

Supplemental Figure 1. To help validate the spectral counting approach for determining protein distribution patterns we compared this method to a quantitative protein measurement technique. Densitometry measurements from a western blot for apoA-I across fractions 18 to 30 (panel B) showed strong distribution similarity to spectral counting data (panel A).

196

Healthy Normalized Spectral Counts A 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 C4b-binding protein alpha chain 0 0 0.5 1 0.323529 0.176471 0.029412 0 0 0 0 0 0 0 0 0 0 0 Apolipoprotein B-100 0 0 0.19375 1 0.73125 0.1375 0 0 0 0 0 0 0 0 0 0 0 0 Fibrinogen alpha chain 0 0 0.005495 0.527473 1 0.791209 0.291209 0.043956 0.016484 0.005495 0 0 0 0.010989 0.005495 0.005495 0.010989 0 Apolipoprotein-L1 0 0 0.285714 0 0 0 1 0.571429 0.142857 0.285714 0.285714 0 0 0 0 0 0 0 Haptoglobin-related protein 0 0 0 0.085714 0 0 0.685714 1 0.771429 0.028571 0 0 0 0 0 0 0 0 AMBP protein 0 0 0 0 0 0 0 0.714286 1 0.142857 0 0 0 0 0 0 0 0.714286 Inter-alpha-trypsin inhibitor heavy chain H2 0 0 0 0 0 0 0 0.298851 1 0.770115 0.425287 0.022989 0 0 0 0 0 0 Apolipoprotein E 0 0 0.5625 0.270833 0 0 0 0.125 0.604167 1 0.895833 0.479167 0.104167 0.020833 0 0.020833 0.041667 0 Plasma protease C1 inhibitor 0 0 0 0 0 0 0 0 0.75 1 0 0 0 0 0 0 0 0 Apolipoprotein C-II 0 0 0.5 0.25 0 0 0 0 0 0.25 1 0.75 0.5 0.25 0 0 0 0 Apolipoprotein C-III 0 0 0.545455 0.909091 0.181818 0.181818 0.363636 0.454545 0.818182 0.727273 1 0.818182 0.272727 0.454545 0.181818 0.090909 0 0 Complement C3 0 0 0 0 0 0 0 0 0.002326 0.276744 1 0.855814 0.527907 0.088372 0.002326 0 0.002326 0 Complement C4-B 0 0 0 0.011321 0 0 0.003774 0.049057 0.139623 0.928302 1 0.373585 0.018868 0 0 0 0 0 Serum amyloid A protein 0 0 0 0 0 0 0 0 0 0.25 1 1 0.5 0 0 0 0 0 Apolipoprotein C-I 0 0 0.307692 0.461538 0 0.076923 0.076923 0.153846 0.307692 0.461538 0.846154 1 0.692308 0.461538 0.615385 0.153846 0.076923 0 Apolipoprotein M 0 0 0 0.333333 0 0 0 0 0.25 0.666667 0.666667 1 0.083333 0 0 0 0 0 Insulin-like growth factor-binding protein ALS 0 0 0 0 0 0 0 0 0 0 0.022222 1 0.888889 0.2 0.022222 0 0 0 Complement C1s subcomponent 0 0 0 0 0 0 0 0 0 0.208333 0.916667 1 0.375 0 0 0 0 0 Inter-alpha-trypsin inhibitor heavy chain H4 0 0 0 0 0 0.009615 0 0 0.009615 0.009615 0.278846 1 0.846154 0.403846 0.057692 0 0.048077 0.019231 Serum amyloid A-4 protein 0 0 0 0 0 0 0 0 0 0.2 0.6 1 0.6 0.2 0 0 0 0 Apolipoprotein A-II 0 0.043478 0 0 0 0 0 0 0.304348 0.565217 0.869565 1 1 0.652174 0.695652 0.086957 0 0 Apolipoprotein A-I 0.024155 0 0.033816 0.10628 0.05314 0.096618 0.120773 0.198068 0.42029 0.594203 0.888889 0.961353 1 0.68599 0.458937 0.371981 0.236715 0.05314 Clusterin 0 0 0 0 0 0.145455 0.218182 0.272727 0.181818 0.163636 0.254545 0.545455 1 0.672727 0.418182 0.036364 0 0 Serum paraoxonase/arylesterase 1 0 0 0 0 0 0 0 0 0.12 0.56 0.92 0.8 1 0.76 0.76 0.48 0.24 0.04 Alpha-1B-glycoprotein 0 0 0 0 0 0 0 0 0 0 0 0 0.21875 1 0.09375 0 0 0 Alpha-2-antiplasmin 0 0 0 0 0 0 0 0 0 0 0 0 0.9 1 0.3 0 0 0 Complement factor B 0 0 0 0 0 0 0 0 0 0 0 0 0.772277 1 0.920792 0.49505 0.009901 0 Hemopexin 0 0 0 0 0 0 0 0 0 0 0 0 0.340426 1 0.93617 0.87234 0.446809 0.021277 Heparin cofactor 2 0 0 0 0 0 0 0 0 0 0 0 0 0.717949 1 0.871795 0.538462 0.128205 0 Kininogen-1 0 0 0 0 0 0 0 0 0.121212 0.393939 0.181818 0.090909 0.424242 1 0.727273 0.575758 0.030303 0 Prothrombin 0 0 0 0 0 0 0 0 0 0 0 0 0.632653 1 0.693878 0.44898 0.020408 0.020408 Serotransferrin 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0.222222 0.166667 0 0 Angiotensinogen 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0.5 0.75 0.75 0 Vitronectin 0 0 0 0 0 0 0 0 0 0.044444 0.044444 0 0.6 1 0.977778 0.555556 0.111111 0 Beta-2-glycoprotein 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0.657143 1 0.6 0.2 0 Alpha-2-HS-glycoprotein 0 0 0 0 0 0 0 0 0 0 0 0 0.037037 0.62963 1 0.888889 0.333333 0 Antithrombin-III 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.477273 1 0.818182 0.068182 Apolipoprotein A-IV 0 0 0 0 0 0 0 0 0.006993 0.125874 0.377622 0.293706 0.076923 0.041958 0.664336 1 0.832168 0.237762 Lipopolysaccharide-binding protein 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0.8 0.2 Serum albumin 0.043478 0.003344 0.010033 0.003344 0 0 0.003344 0.053512 0.140468 0.120401 0.076923 0.053512 0.046823 0.461538 0.862876 1 0.963211 0.448161 Transthyretin 0 0 0 0 0 0 0 0.1 0 0 0 0 0 0.1 0.7 1 1 0.8 Pigment epithelium-derived factor 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.051724 0.793103 1 0.551724 Retinol-binding protein 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.454545 1 0.454545 Alpha-1-antitrypsin 0 0 0 0 0 0 0.1 0.2 0 0 0 0 0 0 0.9 0.8 1 0.3 Vitamin D-binding protein 0 0 0 0 0 0 0 0 0 0 0 0 0.051724 0.051724 0.103448 0.724138 1 0.241379

197

T2DM Normalized Spectral Counts B 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 C4b-binding protein alpha chain 0 0 0.444444 1 0.444444 0 0 0 0 0 0 0 0 0 0 0 0 0 Apolipoprotein B-100 0 0 0.136364 1 0.636364 0 0 0 0 0 0 0 0 0 0 0 0 0 Fibrinogen alpha chain 0 0 0 0.411765 0.901961 1 0.333333 0 0.019608 0 0 0 0 0 0 0 0 0 Apolipoprotein-L1 0 0 0 0 0 0 0.6 1 0.2 0 0.2 0.2 0 0 0 0 0 0 Haptoglobin-related protein 0 0 0 0 0 1 0.6 0.5 0.6 0 0 0 0.1 0 0 0 0 0 AMBP protein 0 0 0 0 0 0 0 0.4 1 0.6 0 0 0 0 0 0 0 0.6 Inter-alpha-trypsin inhibitor heavy chain H2 0 0 0 0 0 0 0 0.5 1 0.5 0.142857 0 0 0 0 0 0 0 Apolipoprotein E 0 0 1 0.25 0 0 0 0 0.041667 0.166667 0.25 0.166667 0.125 0.041667 0.125 0.083333 0.166667 0.041667 Plasma protease C1 inhibitor 0 0 0 0 0 0 0 0 0.5 1 0 0 0 0 0 0 0 0 Apolipoprotein C-II 0 0 1 0 0 0 0 0 0 0.333333 0 0.333333 0 0 0 0 0 0 Apolipoprotein C-III 0 0 0.4 0.4 0.6 0.6 0.4 0.4 0.4 0.8 1 0.8 0.2 0.2 0 0 0.2 0 Complement C3 0 0 0 0 0 0 0 0 0 0.26257 0.910615 1 0.50838 0.01676 0 0 0 0 Complement C4-B 0 0 0 0 0 0 0 0.019802 0.029703 0.841584 1 0.217822 0.009901 0 0 0 0 0 Serum amyloid A protein 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Apolipoprotein C-I 0 0 0.8 0.2 0 0 0 0 0 0 0.2 0.8 1 0.4 0.2 0 0 0 Apolipoprotein M 0 0 0 0 0 0 0 0 0 0 0.5 1 0.25 0 0 0 0 0 Insulin-like growth factor-binding protein ALS 0 0 0 0 0 0 0 0 0 0 0 1 0.923077 0.076923 0 0 0 0 Complement C1s subcomponent 0 0 0 0 0 0 0 0 0 0 0.3 1 0.1 0 0 0 0 0 Inter-alpha-trypsin inhibitor heavy chain H4 0 0 0 0 0 0 0 0 0 0 0.086957 1 0.869565 0.130435 0 0 0 0 Serum amyloid A-4 protein 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 Apolipoprotein A-II 0 0 0 0 0 0 0 0.090909 0 0 0.363636 1 0.636364 0.727273 0.454545 0.181818 0 0 Apolipoprotein A-I 0.053763 0 0 0.043011 0.010753 0.032258 0.204301 0.172043 0.225806 0.365591 0.516129 0.784946 1 0.677419 0.602151 0.451613 0.311828 0.107527 Clusterin 0 0 0 0 0 0 0.217391 0.130435 0.217391 0.26087 0.26087 0.826087 1 0.782609 0.391304 0 0 0 Serum paraoxonase/arylesterase 1 0 0 0 0 0 0 0 0 0 0.2 0 0.6 1 0.2 0.8 0.4 0 0 Alpha-1B-glycoprotein 0 0 0 0 0 0 0 0 0 0 0 0 0.111111 1 0.111111 0 0 0 Alpha-2-antiplasmin 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 Complement factor B 0 0 0 0 0 0 0 0 0 0 0 0 0.478261 0.826087 1 0.23913 0 0 Hemopexin 0 0 0 0 0 0 0 0 0 0 0 0 0.318182 0.818182 1 0.727273 0.181818 0 Heparin cofactor 2 0 0 0 0 0 0 0 0 0 0 0 0 0.565217 1 0.73913 0.521739 0 0 Kininogen-1 0 0 0 0 0 0 0 0 0.125 0.125 0.0625 0.0625 0.4375 0.625 1 0.4375 0 0 Prothrombin 0 0 0 0 0 0 0 0 0 0 0 0 0.85 1 0.65 0.2 0 0 Serotransferrin 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Angiotensinogen 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 0 Vitronectin 0 0 0 0 0 0 0 0 0 0 0 0 0.458333 1 0.875 0.583333 0 0 Beta-2-glycoprotein 1 0 0 0 0 0 0 0 0 0 0 0 0 0.076923 0.538462 1 0.538462 0.076923 0 Alpha-2-HS-glycoprotein 0 0 0 0 0 0 0 0 0 0 0 0 0 0.888889 1 1 0.222222 0 Antithrombin-III 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.3125 1 0.9375 0.0625 Apolipoprotein A-IV 0 0 0 0 0 0 0 0 0 0 0.25 0.211538 0 0.019231 0.634615 1 0.788462 0.211538 Lipopolysaccharide-binding protein 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0.5 0 Serum albumin 0 0 0 0 0 0 0 0.017857 0.053571 0.080357 0.035714 0.053571 0 0.321429 0.866071 1 0.964286 0.375 Transthyretin 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.5 1 1 1 Pigment epithelium-derived factor 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.428571 1 0.321429 Retinol-binding protein 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 Alpha-1-antitrypsin 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0.5 1 0 Vitamin D-binding protein 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.090909 0.454545 1 0.181818

198

Chapter 7. Discussion

At the start of work on this thesis we developed a hypothesis that was based on years of cumulative research dating back to the 1950’s when HDL was first isolated. This research included numerous epidemiological studies unanimously demonstrating the existence of a strong inverse relationship between circulating levels of HDL-C and the incidence of various manifestations of cardiovascular disease 1, 2. This newfound importance of HDL drove interest in detailed studies determined to identify the mechanism behind this protection. After more than 60 years, the results of these studies point to a complexity of HDL that is far beyond what anybody could have imagined. This class of lipoprotein has been demonstrated to possess a staggering number of direct mechanistic functions with diverse physiological roles ranging from lipid metabolism to innate immune defense 3. This functional diversity of HDL may be partially explained by studies of composition. Biochemical analyses of total HDL isolated from human plasma have shown a variety of lipid and protein components 4, 5, which may even change in response to changes in the local environment of the plasma

(e.g. inflammation, disease). Recent applications of high resolution mass spectrometry to examine the HDL proteome further support this functional diversity by consistently identifying around 55 proteins in association with total HDL (Table 1-1). Based on this existing field of knowledge we state our hypothesis: The total pool of HDL in an individual is composed of numerous compositionally distinct subspecies with varying functionality, and these subspecies are likely altered in disease. The aims of this thesis

199 seek to take this field a step further, by mending together studies of HDL function and composition with the ultimate goal of pushing the field toward a new classification scheme for HDL that is based on particle functionality and composition rather than some arbitrary physicochemical property inherent to the isolation technique used to collect it, as done in the past.

Thesis summary

Our studies begin with a complete proteomic characterization of HDL subfractions. In

Chapter 2 we describe a novel method which we have developed for the analysis of

HDL subfractions isolated by gel filtration chromatography. This new technique utilizes a phospholipid binding resin (LRA) to isolate only those proteins which associate with lipid. This represents an important advance for the field as it allows for the analysis of lipoproteins separated by techniques that are less physically disruptive to the particle structure and composition than ultracentrifugation. Prior to this advance, proteomic analysis of HDL isolated by methods that separate by size or physical properties other than density was not possible due to the co-migration of lipid free proteins with lipoproteins. It would be difficult to tell if identified proteins are in fact associated with a lipid particle, so by isolating only those proteins which associate with phospholipid we can identify lipoprotein associated proteins. Additionally, we have developed a new high resolution gel filtration chromatography method which provides a higher degree of fractionation of total HDL than previously described gel filtration techniques. This is very important to the aims of our thesis as we are interested in separating out and

200 analyzing as many of these HDL subpopulations as possible. Using these new techniques we analyzed the HDL we can tell that the fractions collected likely contain multiple subspecies of HDL, i.e. we did not get separation of individual subspecies preventing us from determining exact protein composition of individual particle species.

To overcome this obstacle we developed a novel multifaceted correlation strategy (Chapter 3). This strategy utilized protein distribution data from multiple separation techniques to identify pairs of proteins which are likely to reside together on a single HDL particle. Three separation techniques were used with each separating

HDL into multiple subfractions based on a different physical property. The idea is that each technique with spread the total HDL out differently, but those proteins that are physically constricted to a given particle together will consistently be found in a fraction together regardless of the separation technique used. Our analysis statistically combined protein correlation networks from each separation technique to produce a list ranking pairs of proteins based of HDL subspecies (Chapter 4). To do this we started by using mouse knockout models for the proteins apoA-I, apoA-II and apoA-IV. The protein distribution patterns across gel filtration fractions from each mouse type was compared to wild type mice with a statistical analysis designed to determine shifts in protein distribution. A shift in distribution of a given ~65% of total protein but only affected the distribution patterns for very few minor proteins. To supplement these studies, we were lucky to find a human patient with a genetic deficiency of apoA-I and who was willing to participate in our study. In the human system, it appeared that the absence of apoA-I affected the

201 distributions of a greater percentage of the minor proteins, suggesting that the role of this protein in HDL subspecies formation in humans may be greater than in the mouse.

Experiments related to Aim 2 were focused on the functional analysis of HDL subfractions (Chapter 5). Here we performed a panel of assays for the best known HDL functions on gel filtration fractions from the individuals whose proteome was analyzed in

Chapter 2. We measured activity of individual fractions for cholesterol efflux, inhibition of LDL oxidation and anti-inflammatory capacity. In these studies we found that activity for each function was distributed differently across fractions supporting our hypothesis that particles of different composition would possess different functionality. We then combined this functional data with our proteomics data in a functional correlation analysis that allowed us to identify protein components of HDL that are likely driving these functions.

The goal of Aim 3 was to examine the influence of a specific metabolic condition, type 2 diabetes, on the composition of HDL subspecies. We obtained funding for a relatively small pilot study on a population of type 2 diabetics and control participants and compared distribution profiles for lipids and proteins across gel filtration fractions, using the methods we developed in Chapter 2. We found that in T2D, participants had significantly decreased levels of phospholipid associated with fractions containing large

HDL and that this decrease correlated very strongly with vascular health as measured by pulse wave velocity. Additionally, proteomics data indicated a corresponding decrease in several HDL associated proteins in these fractions with known functions in lipid transport and anti-oxidation. This data suggests that there may be a specific subspecies of large HDL that becomes dysfunctional or is absent in patients with T2D

202 and may be the cause of the increased rate of atherosclerosis observed in these patients.

Contributions of this work

Understanding HDL composition. Taken together the studies presented here represent several important advances in the study of HDL subspecies. We have developed new methods for the isolation of HDL subspecies and have described the composition and function of HDL subfractions and how these correlate with each other resulting in the existence of subpopulations of HDL with varied composition and function, directly supporting our global hypothesis. Based on the composition experiments and the protein interaction experiments in the various apolipoprotein deficient model systems, we believe that proteins are associating with HDL in a very dynamic manner. Different proteins likely associate with HDL by different mechanisms; some having direct lipid binding affinities and other associating via protein to protein interactions with other proteins which are capable of binding lipid. It seems likely that proteins that associate with HDL can be divided into two groups: strict lipoprotein associated proteins and exchangeable proteins (Figure 7-1). Where strict lipoprotein associated proteins are almost always bound to a lipoprotein and are a structural component providing stability to the particle and exchangeable proteins associate peripherally and more weakly such that they can pop on an off of the particle with little energy. Additionally, we think that the available proteins in these groups can be altered by physiological changes in the body that may alter the availability of these proteins in

203 the plasma such as general inflammation or some specific disease. Our studies of HDL subfractions in patients with T2D support this by showing that subspecies composition can be altered by disease and this potentially is responsible for clinical consequences.

We strongly believe that this may be the case for other types of disease as well and that this area deserves further investigation.

Redefining “HDL”. Our studies in the apoA-I knockout mouse model and in the apoA-I deficient human suggest that although many think of HDL as being defined by this protein, there still exist lipidated particles in the size range of traditional HDL even in its absence. This tells us that perhaps this apoA-I centric view of HDL is incorrect and although a large majority of circulating HDL do in fact contain apoA-I, this protein is not an absolute requirement for the formation of HDL, it may be the case that some of these minor HDL associated proteins are capable of forming HDL particles on their own.

From the viewpoint of understanding the function of individual particles within the

HDL milieu, our current definitions of HDL may fall short. So what is the best way to define HDL? As we alluded to before, it may be that a more holistic view of HDL is warranted, one based on composition rather than method of separation. One universal constituent of all HDL subfractions is phospholipid. In cells, a critical role of the plasma membrane is to act as a two-dimensional solvent for the concentration of phospholipid- bound proteins and assembly into productive complexes. Perhaps phospholipids play an analogous role in circulation by acting as an organizing center for the assembly of extracellular lipid-associating proteins. When most small exchangeable amphipathic proteins interact with phospholipids, the resulting complexes usually fall into the size

204 and density range of classically defined HDL 6. Thus, HDL probably contains a host of protein/lipid complexes of defined composition that are as yet uncharacterized (and may not contain apoA-I at all) but may still carry out important biological functions. Given this, it may be useful to think of HDL in terms of “phospholipid-rich lipoproteins” (PL-rich

LPs). This definition distinguishes from chylomicrons/VLDL (triglyceride-rich lipoproteins) and LDL (cholesteryl ester rich) and frees them from traditional apoA-I centric or density-centric preconceptions that accompany the term ‘HDL’ (Figure 7-2).

Circulating phospholipid may be the platform upon which specific protein-protein or protein-lipid interactions drive the formation of distinct particles with distinct functions.

Taking into account the diverse set of functions and adding in the increasing appreciation of the compositional and structural heterogeneity of HDL, it is difficult to imagine that all these functions are mediated by the relatively limited number of HDL subspecies that are currently characterized. It is becoming clear that the term “HDL” refers to an ensemble of discrete particles - each with their own complement of proteins and lipids that endow the host particle with distinct and quite varied functionalities.

Indeed, many of these particles may play important physiological roles that have little to do with RCT or protection from heart disease. Some of these functions are closely related to the ability of HDL to modify the behavior of a target cell or organism by removing lipids. However, an increasing number appear to be mediated via interactions with cell surface proteins to trigger distinct signaling pathways to alter cell function 3.

While many of these studies are still in the beginning stages, it would seem that it is an exciting time to be studying HDL metabolism.

205

HDL subspecies in cardiovascular disease. Although a large body of epidemiological evidence suggests the measurement of HDL-C as a biomarker for CVD, there still exist many individuals with optimal HDL-C levels who still develop CVD and vice versa. HDL-C is only one measure, that of the amount of cholesterol currently carried in HDL particles, perhaps another measure of HDL could do better at predicting this risk, such as another compositional component like phospholipid or something more directly related to the protective functionality of HDL. With our new ability to examine subspecies of HDL more closely we can begin to evaluate this possibility. Our functional data in Chapter 5 suggest that phospholipid is a driving force for cholesterol efflux so perhaps phospholipid content of HDL would better describe its capacity for cholesterol efflux, and therefore cardio-protection, in vivo. Our studies in Aim 3 have identified a size fraction of HDL (F22) that is altered in type 2 diabetes and strongly correlates with cardiovascular health. The lab is currently designing an expanded study to further characterize this HDL particle and explore its potential in predicting cardiovascular health. We may even aim to explore strategies for raising this HDL species specifically with the hope of reducing cardiovascular disease. Using the tools developed here we also may be able to evaluate the impact of currently existing therapeutic strategies the HDL subspecies profile. The methods described in this thesis are great for discovery purposes however will likely have little direct clinical utility as they stand because they are quite time consuming and require expensive equipment.

This would not work well as a clinical assay, however as we continue to characterize these HDL subspecies and determine which are playing the most important roles in the body we can begin to develop assays for direct measurement of individual subspecies.

206

One possibility would be to raise antibodies that specifically recognized the conformation of a cluster of proteins on an HDL species so that we could develop an

ELISA style assay system.

The future of HDL subspecies research and clinical implications

In our view, two key challenges lie before the field of HDL subspecies research. First, we need a better understanding of the subparticle makeup of the fractions classically referred to as “HDL”. New technologies for alternative particle isolation and analysis and clever strategies for identifying distinct particle functions on the background of staggering compositional complexity will have to be developed to meet this important challenge. Once these subspecies are identified and characterized, it will be easier to correlate specific functions for these particles. Additionally, further in vitro experimentation needs to be done to better understand the functional implications of the different subspecies and in vivo animal experiments as well to identify mechanisms for manipulating individual subspecies and the clinical outcomes. Second, research needs to be directed at identifying additional roles for HDL outside of the classical purview of cardiovascular disease. The strong tie in with inflammation and innate immunity makes

HDL a promising target for treatment or prevention of diseases such as those caused by opportunistic pathogens or chronic inflammatory states. This appreciation will not only be useful for understanding the potential of HDL as a treatment option for non- cardiovascular pathologies, but it may be also be critical for understanding the consequences of pharmacological manipulation of HDL-C levels. Indeed, current

207 pharmacological therapies such as niacin or those in development including CETP inhibition and apoA-I transcription stimulation aim to raise plasma HDL cholesterol in the generic sense without direct knowledge of the functionality (or lack thereof) of the elevated species. It will be important to ascertain the impact of these manipulations, not only on those subspecies linked to CVD protection, but also on these peripheral functions that may be just as important.

The advancements presented in this thesis will provide data and tools that will be useful as we develop exciting new approaches for the development of HDL as a biomarker and therapeutic target for a wide variety of disease states.

208

Reference List

1 BARR DP, RUSS EM, EDER HA. Protein-lipid relationships in human plasma. II. In

atherosclerosis and related conditions. Am J Med 1951;11(4):480-93.

2 Gordon T, Castelli WP, Hjortland MC, Kannel WB, Dawber TR. High density

lipoprotein as a protective factor against coronary heart disease. The Framingham

Study. Am J Med 1977;62(5):707-14.

3 Gordon SM, Hofmann S, Askew DS, Davidson WS. High density lipoprotein: it's not

just about lipid transport anymore. Trends Endocrinol Metab 2011;22(1):9-15.

4 Vaisar T, Mayer P, Nilsson E, Zhao XQ, Knopp R, Prazen BJ. HDL in humans with

cardiovascular disease exhibits a proteomic signature. Clin Chim Acta 2010;411(13-

14):972-9.

5 Vaisar T, Pennathur S, Green PS, Gharib SA, Hoofnagle AN, Cheung MC, Byun J,

Vuletic S, Kassim S, Singh P, Chea H, Knopp RH, Brunzell J, Geary R, Chait A,

Zhao XQ, Elkon K, Marcovina S, Ridker P, Oram JF, Heinecke JW. Shotgun

proteomics implicates protease inhibition and complement activation in the

antiinflammatory properties of HDL. J Clin Invest 2007;117(3):746-56.

6 Davidson WS, Jonas A, Clayton DF, George JM. Stabilization of alpha-synuclein

secondary structure upon binding to synthetic membranes. J Biol Chem

1998;273(16):9443-9.

209

Figure 7-1. Model for association of protein components with HDL.

210

VLDL LDL HDL

Component VLDL LDL HDL (% mass) Protein 7-12 21 30-60 Phospholipid 18 20 30 Cholesterol (total) 20 50 17 Triglyceride 50 4 3

Figure 7-2. Defining lipoproteins based on primary lipid components.

211